ROUTE OPTIMIZATION

Prepared by:
William Schneider
Tyler M. Miller
William A. Holik

Prepared for:
The Ohio Department of Transportation,
Office of Statewide Planning & Research

State Job Number: 135157

November 2016
Final Report
 Technical Report Documentation Page

1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

FHWA/OH-2016-20

4. Title and Subtitle 5. Report Date

November, 2016
Route Optimization Phase 1 6. Performing Organization Code

7. Author(s) 8. Performing Organization Report No.

William Schneider, Tyler M. Miller, and William A. Holik
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
The University of Akron
302 Buchtel Common 11. Contract or Grant No.
Akron, Ohio 44325-2102 27093
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
Ohio Department of Transportation Draft Report
1980 West Broad Street 14. Sponsoring Agency Code
Columbus, Ohio 43223
15. Supplementary Notes
Project performed in cooperation with the Ohio Department of Transportation.
16. Abstract
For winter maintenance purposes, the Ohio Department of Transportation (ODOT) deploys a fleet of
approximately 1,600 snow plow trucks that maintain 43,000 lane miles of roadway. These trucks are based out
of 200 garages, yards, and outposts that also house 650,000 tons of salt (ODOT, 2011). The deployment of
such a large number of trucks over a vast maintenance area creates an operational problem in determining the
optimal maintenance routes and fleet size. In recent years, several advances have been made in route
optimization that may aid in determining the required number of trucks and the area that these trucks should
maintain throughout the state of Ohio. Traditionally, ODOT has used county borders as maintenance
boundaries for ODOT garages. However, by removing these borders and optimizing the snow plow routes,
ODOT may benefit from a significant time and cost savings. For the purpose of route optimization, ODOT
Districts 1, 2, and 10 have been selected to serve as case studies for this project.

The results of this project will provide ODOT a tool to determine the minimum number of trucks needed to
maintain the necessary roadways within Districts 1, 2, and 10. In addition, the project provides ODOT a tool to
assign assets to specific facilities and the most optimal routes for each truck in the district. This research may
result in a reduction of fleet sizes and a significant cost savings while maintaining an equal or better LOS.

17. Keywords 18. Distribution Statement

Route Optimization, Level of Service, Winter Maintenance, Fleet
Optimization
19. Security Classification (of 20. Security Classification
this report) (of this page) 21. No. of Pages 22. Price
Unclassified Unclassified 102

Form DOT F 1700.7 (8-72) Reproduction of completed pages authorized

Final Report i
 ROUTE OPTIMIZATION

Prepared by:

William Schneider, Ph.D., P.E.

Tyler Miller,
William A. Holik, Ph.D.
Department of Civil Engineering
The University of Akron

November 2016

Prepared in cooperation with the Ohio Department of Transportation

The contents of this report reflect the views of the author(s) who is (are) responsible for the facts and the accuracy
of the data presented herein. The contents do not necessarily reflect the official views or policies of the Ohio
Department of Transportation or the Federal Highway Administration. This report does not constitute a standard,
specification, or regulation.

Final Report ii
 ACKNOWLEDGMENTS

This project was conducted in cooperation with Ohio Department of Transportation (ODOT).

The authors would like to thank the members of ODOT’s Technical Liaison Committee:

 Jamie Hendershot,
 Rod Nuveman,
 Hiram Crabtree,
 Layth Istefan, and
 Fred Judson.
The time and input provided for this project by Technical Liaison Committee are greatly appreciated. In
addition to our technical liaisons, the authors would like to express their appreciation to Ms. Michelle
Lucas, Ms. Jill Martindale, Ms. Cynthia Jones, Mr. Scott Phinney and Ms. Kelly Nye from ODOT’s
Office of Statewide Planning & Research for their time and assistance. The authors would also like to
thank Benjamin Gleichert, Salvatore Valeriano, Zachery Teter, Zachary Gould, Matthew Devlin, Ryan
Bonzo, and Jamie Lowery of The University of Akron for their help on this project.

Final Report iii

 Customary Customary
SI Unit Factor SI Unit Factor
Unit Unit
Length Length
inches millimeters 25.4 millimeters inches 0.039
inches centimeters 2.54 centimeters inches 0.394
feet meters 0.305 meters feet 3.281
yards meters 0.914 meters yards 1.094
miles kilometers 1.61 kilometers miles 0.621
Area Area
square square
square inches 645.1 square inches 0.00155
millimeters millimeters
square square
square feet 0.093 square feet 10.764
meters meters
square square
square yards 0.836 square yards 1.196
meters meters
acres hectares 0.405 hectares acres 2.471
square square
square miles 2.59 square miles 0.386
kilometers kilometers
Volume Volume
gallons liters 3.785 liters gallons 0.264
cubic feet cubic meters 0.028 cubic meters cubic feet 35.314
cubic yards cubic meters 0.765 cubic meters cubic yards 1.308
Mass Mass
ounces grams 28.35 grams ounces 0.035
pounds kilograms 0.454 kilograms pounds 2.205
short tons megagrams 0.907 megagrams short tons 1.102

Final Report iv
 TABLE OF CONTENTS

CHAPTER I INTRODUCTION ............................................................................................................ 1

1.1 Purposes and Objectives ............................................................................................................... 1

1.2 Benefits from this Research .......................................................................................................... 1

1.3 Organization of this Report ........................................................................................................... 2

CHAPTER II BACKGROUND .............................................................................................................. 3

2.1 Literature Review.......................................................................................................................... 3

2.2 Route Optimization Tools ............................................................................................................. 5

2.2.1 Network Analyst ................................................................................................................... 5

2.2.2 Vehicle Routing Problem ...................................................................................................... 5

2.2.3 QTravel ................................................................................................................................. 6

CHAPTER III PROJECT SETTING ........................................................................................................ 7

3.1 Project Setting ............................................................................................................................... 7

3.2 ODOT District 1 ........................................................................................................................... 9

3.3 ODOT District 2 ......................................................................................................................... 11

3.3.1 District 2 Part 1 ................................................................................................................... 12

3.3.2 District 2 Part 2 ................................................................................................................... 13

3.3.3 District 2 Part 3 ................................................................................................................... 15

3.4 ODOT District 10 ....................................................................................................................... 17

CHAPTER IV ROUTE OPTIMIZATION METHODOLOGY .............................................................. 19

4.1 Initial Model Creation ................................................................................................................. 19

4.2 Data Collection ........................................................................................................................... 21

4.3 Creating the Route Optimization Model ..................................................................................... 22

4.3.1 Loading Plowing Locations ................................................................................................ 22

4.3.2 Route Restrictions ............................................................................................................... 24

4.3.3 Assigning Cycle Times ....................................................................................................... 25

4.4 Initial Route Optimization .......................................................................................................... 26

Final Report v
 4.5 Verification Process .................................................................................................................... 27

4.6 Fleet Optimization Methodology ................................................................................................ 28

CHAPTER V CURRENT ROUTE ANALYSIS ................................................................................... 29

5.1 District 1 Current Routes Overview............................................................................................ 30

5.1.1 District 1 Current Route Analysis ....................................................................................... 31

5.2 District 2 Current Route Overview ............................................................................................. 32

5.2.1 District 2 Current Route Analysis ....................................................................................... 33

5.3 District 10 Current Route Overview ........................................................................................... 35

5.3.1 District 10 Current Route Analysis ..................................................................................... 36

5.4 Current Route Analysis Summary .............................................................................................. 37

CHAPTER VI INITIAL ROUTE OPTIMIZATION .............................................................................. 38

CHAPTER VII ROUTE VERIFICATION .......................................................................................... 44

7.1 District 1 Route Verification Overview ...................................................................................... 46

7.2 District 1 Route Verification Results .......................................................................................... 46

7.3 District 2 Route Verification Overview ...................................................................................... 48

7.4 District 2 Route Verification Results .......................................................................................... 49

7.4.1 District 2 Part 1 Route Verification Results ........................................................................ 49

7.4.2 District 2 Part 2 Route Verification Results ........................................................................ 51

7.4.3 District 2 Part 3 Route Verification Results ........................................................................ 53

7.5 District 10 Route Verification Overview .................................................................................... 55

7.6 District 10 Route Verification Results ........................................................................................ 56

CHAPTER VIII FLEET OPTIMIZATION ........................................................................................... 59

8.1 District 1 Fleet Optimization Results .......................................................................................... 63

8.1.1 District 1 Alternative Fleet Sizes ........................................................................................ 67

8.1.2 District 1 Fleet Optimization Results Summary ................................................................. 68

8.2 District 2 Fleet Optimization Results .......................................................................................... 70

8.2.1 District 2 Part 1 Fleet Optimization Results ....................................................................... 70

Final Report vi
 8.2.2 District 2 Part 2 Fleet Optimization Results ....................................................................... 72

8.2.3 District 2 Part 3 Fleet Optimization Results ....................................................................... 75

8.3 District 10 Fleet Optimization Results ........................................................................................ 78

CHAPTER IX ROUTE VULNERABILITY .......................................................................................... 83

9.1 District 1 Route Vulnerability ..................................................................................................... 83

9.2 District 2 Route Vulnerability ..................................................................................................... 84

9.2.1 District 2 Part 1 Vulnerability ............................................................................................. 84

9.2.2 District 2 Part 2 Vulnerability ............................................................................................. 85

9.2.3 District 2 Part 3 Vulnerability ............................................................................................. 86

9.3 District 10 Route Vulnerability ................................................................................................... 86

CHAPTER X IMPLEMENTATION ..................................................................................................... 88

10.1 Recommendations for the Implementation of the Optimized Routes ......................................... 88

10.2 Steps Needed to Implement Findings ......................................................................................... 88

10.3 Suggested Time Frame for Implementation ................................................................................ 89

10.4 Benefits Expected from Implementation .................................................................................... 89

10.5 Potential Risks and Obstacles to Implementation ....................................................................... 89

10.6 Strategies to Overcome Potential Risks and Obstacles ............................................................... 90

10.7 Potential Users and Other Organizations .................................................................................... 90

10.8 Estimated Cost of Implementation .............................................................................................. 90

REFERENCES ........................................................................................................................................... 91

APPENDIX A: INITIAL OPTIMIZED DISTRICT MAPS ....................................................................... 93

APPENDIX B: FLEET OPTIMIZED DISTRICT MAPS .......................................................................... 98

Final Report vii

 LIST OF FIGURES
Figure 3.1: ODOT Districts involved with the Route Optimization Project. ................................................ 7

Figure 3.2: Average Yearly Snowfall in Ohio. ............................................................................................. 8

Figure 3.3: District 1 Elevation. .................................................................................................................... 9

Figure 3.4: Example of Snow Blowing onto Road. .................................................................................... 10

Figure 3.5: District 1 Routes and Facility Locations. ................................................................................. 10

Figure 3.6: District 2 Elevation. .................................................................................................................. 11

Figure 3.7: District 2 Part 1 Route Optimization Facility Locations. ......................................................... 12

Figure 3.8: District 2 Part 2 Route Optimization Facility Locations. ......................................................... 13

Figure 3.9: Additional Outpost in Wood County. ....................................................................................... 14

Figure 3.10: Change in Sandusky County Garage Location. ...................................................................... 15

Figure 3.11: District 2 Part 3 Route Optimization Facility Locations. ....................................................... 16

Figure 3.12: District 2 Part 3 Wood County Garage Change in Location. ................................................. 16

Figure 3.13: District 10 Elevation. .............................................................................................................. 17

Figure 3.14: District 10 Routes and Facility Locations. ............................................................................. 18

Figure 4.1: Initial Route Optimization Model Development. ..................................................................... 19

Figure 4.2: Example of Elevation Differences and Directionality for Roads. ............................................ 20

Figure 4.3: ODOT Snow and Ice Routes. ................................................................................................... 21

Figure 4.4: Snow and Ice Routes within District 10. .................................................................................. 23

Figure 4.5: Plowing Locations within District 10 ....................................................................................... 24

Figure 4.6: Example of Route Restriction Implementation in District 10. ................................................. 25

Figure 4.7: The GPS Transponder for Collecting Data from Driving the Proposed Routes. ...................... 27

Figure 5.1: Current Route Analysis Methodology. ..................................................................................... 29

Figure 5.2: District 1 Current Route Overview........................................................................................... 30

Figure 5.3: District 2 Current Treating Areas by Facility. .......................................................................... 33

Figure 5.4: District 10 Current Treating Areas by Facility. ........................................................................ 35

Figure 6.1: Initial Route Optimization Methodology. ................................................................................ 38

Final Report viii

Figure 6.2: Example of District 2 Part 1 Overview Map. ........................................................................... 39

Figure 6.3: Example of a Map produced from the Initial Optimized Routes in Fulton County. ................ 40

Figure 6.4: Example of the Route Descriptions used to accompany the Facility Maps.............................. 41

Figure 6.5: Example of Table and Graphs Produced to Show Number of Cycles Before Refill. ............... 42

Figure 7.1: Route Verification Process. ...................................................................................................... 44

Figure 7.2: Map of Verification Plan for ODOT District 1. ....................................................................... 46

Figure 7.3: District 1 Initial Route Verification Data. ................................................................................ 47

Figure 7.4: District 1 Reconfigured Verification Data................................................................................ 48

Figure 7.5: Map of Verification Plan for ODOT District 2. ....................................................................... 49

Figure 7.6: District 2 Part 1 Initial Route Verification Data. ...................................................................... 50

Figure 7.7: District 2 Part 1 Reconfigured Verification Data. .................................................................... 51

Figure 7.8: District 2 Part 2 Initial Route Verification Data. ...................................................................... 52

Figure 7.9: District 2 Part 2 Reconfigured Verification Data. .................................................................... 53

Figure 7.10: District 2 Part 3 Initial Route Verification Data. .................................................................... 54

Figure 7.11: District 2 Part 3 Reconfigured Verification Data. .................................................................. 55

Figure 7.12: Map of Verification Plan for ODOT District 10. ................................................................... 56

Figure 7.13: District 10 Initial Route Verification Data. ............................................................................ 57

Figure 7.14: District 10 Reconfigured Verification Data............................................................................ 58

Figure 8.1: Fleet Optimization Methodology. ............................................................................................ 59

Figure 8.2: Example of the Minimum Truck Crew for Hardin County Garage.......................................... 61

Figure 8.3: Example of Hardin County Garage Secondary Routes with One Truck Removed. ................. 62

Figure 8.4: District 1 LOS Analysis............................................................................................................ 65

Figure 8.5: Fleet Optimized Scenarios. ....................................................................................................... 68

Figure 8.6: District 2 Part 1 LOS Analysis. ................................................................................................ 71

Figure 8.7: District 2 Part 2 LOS Analysis. ................................................................................................ 74

Figure 8.8: District 2 Part 3 LOS Analysis. ................................................................................................ 77

Figure 8.9: District 10 LOS Analysis.......................................................................................................... 80

Final Report ix
Figure 9.1: District 1 Demanding Areas. .................................................................................................... 84

Figure 9.2: District 2 Part 1 Demanding Areas. .......................................................................................... 85

Figure 9.3: District 2 Part 2 Demanding Areas. .......................................................................................... 85

Figure 9.4: District 2 Part 3 Demanding Areas. .......................................................................................... 86

Figure 9.5: District 10 Demanding Areas. .................................................................................................. 87

Final Report x
 LIST OF TABLES

Table 2.1: Literature Review Summary ........................................................................................................ 3

Table 4.2: Level of Service within ODOT Districts 1, 2, and 10. .............................................................. 26

Table 5.1: District 1 Current Route Analysis. ............................................................................................. 32

Table 5.2: District 2 Current Route Analysis. ............................................................................................. 34

Table 5.3: District 10 Current Route Analysis. ........................................................................................... 37

Table 8.1: District 1 Route Optimization Summary of Results. ................................................................. 63

Table 8.2: Recommended Operational Trucks at Each Facility within District 1. ..................................... 66

Table 8.3: Level of Service Increase if Current Operational Trucks are Optimized................................... 67

Table 8.4: District 1 Maximum Level of Service Attainment. ................................................................... 68

Table 8.5: District 1 Operational Truck Assignments for Desired Outcomes. ........................................... 69

Table 8.6: District 2 Part 1 Fleet Optimization Results. ............................................................................. 70

Table 8.7: District 2 Part 1 Operational Truck Assignments. ..................................................................... 72

Table 8.8: District 2 Part 2 Fleet Optimization Results. ............................................................................. 73

Table 8.9: District 2 Part 2 Operational Truck Assignments. ..................................................................... 75

Table 8.10: District 2 Part 3 Fleet Optimization Results. ........................................................................... 76

Table 8.11: District 2 Part 3 Operational Truck Assignments. ................................................................... 78

Table 8.12: District 10 Route Optimization Summary of Results. ............................................................. 79

Table 8.13: Recommended Operational Trucks at Each Facility in District 10. ........................................ 81

Final Report xi
 LIST OF ACRONYMS

lbs/ln mile Pounds per Lane Mile

ODOT Ohio Department of Transportation

GIS Geographic Information System

GPS Global Positioning System

VRP Vehicle Routing Problem

ROM Route Optimization Model

LOS Level of Service

TIMS Transportation Information Mapping System

Final Report xii

Final Report xiii
 CHAPTER I INTRODUCTION

For winter maintenance purposes, the Ohio Department of Transportation (ODOT) deploys a fleet of
approximately 1,600 snow plow trucks that maintain 43,000 lane miles of roadway. These trucks are
based out of 200 garages, yards, and outposts that also house 650,000 tons of salt ( The Ohio Department
of Transportation, 2011). The deployment of such a large number of trucks over a vast maintenance area
creates an operational problem in determining the optimal maintenance routes and fleet size. In recent
years, several advances have been made in route optimization that may aid in determining the required
number of trucks and the area that these trucks should maintain throughout the state of Ohio.
Traditionally, ODOT has used county borders as maintenance boundaries for ODOT garages. However,
by removing these borders and optimizing the snow plow routes, ODOT may realize a significant time
and cost savings.

1.1 Purposes and Objectives

The purposes of this project are to optimize snow and ice routes for ODOT’s snow plow trucks in
Districts 1, 2, and 10 while eliminating county border restrictions. In order to ensure that the purposes of
this project were satisfied, the University of Akron research team developed the following objectives:

 Objective One – Digitize base routes and input ODOT facilities and plowing locations;

 Objective Two – Remove county border restrictions and optimize routes for each truck;

 Objective Three – Place GPS recorders in trucks and collect data regarding actual cycle
times; and

 Objective Four – Set maximum cycle times and determine which garages may remove trucks
and which need additional trucks.

1.2 Benefits from this Research

There are numerous benefits expected from the outcome of this project. One important benefit will be an
analysis that justifies the fleet size in three of ODOT’s twelve districts and, accordingly ensures that
ODOT maintains all of the required roadways within the involved districts in an efficient and economical
manner. In addition, each facility within Districts 1, 2, and 10 will know the specific roadways that it
must maintain, regardless of the amount of resources available during winter maintenance operations.
Another benefit of this research is that the Route Optimization Model (ROM) serves as a tool to analyze
new equipment technology and new operational considerations. This tool is invaluable for ODOT district
leadership as they determine where to allocate limited resources within the district. This benefit extends

Final Report 1
further by revealing areas of concern within the district, thus guiding future facility location and
construction.

1.3 Organization of this Report

This report is divided into ten chapters. Chapter 1 introduces the topic and defines the objectives to be
completed for this research project. Chapter 2 provides background information obtained prior to the
beginning of the project as well as the tools that were utilized. Chapter 3 provides information of the
project setting, in particular, the districts involved and their characteristics. Chapter 4 presents the
methodology of the route optimization. Chapter 5 consists of an analysis of the current routes being used
for winter maintenance operations. Chapter 6 presents the results obtained from the initial optimized
routes. Chapter 7 provides the route verification process and results for each district. Chapter 8
summarizes the fleet optimization for each district. Chapter 9 presents the vulnerable areas for each
district. Chapter 10 presents the implementation plan for the optimized trucks within each district’s fleet.

Final Report 2
 CHAPTER II BACKGROUND

This chapter provides information regarding the background of the project to include a literature
review and route optimization tools used for the project. This chapter is divided into two sections:

 Section One – Literature review; and

 Section Two – Route optimization tools.

2.1 Literature Review

Upon determining the potential cost savings regarding the optimization of winter maintenance operations,
the research team conducting a literature review on how optimization models were created and
implemented within other agencies and organizations. The literature review consists of articles published
academically and a look at state Departments of Transportation regarding the optimization of fleets or
individual vehicles. A summary of the findings is shown in Table 2.1 below.

Table 2.1: Literature Review Summary

Project Goal Methodology Findings Reference

Provide a sustainable
Develop a model to take The analysis shows that
optimization of winter
route optimization away current maintenance
maintenance service by
from human managers fleet numbers and depot (Gudac, Hanak, &
maximizing the potential of
and use computer models locations are sufficient Marovic, 2014)
the service to ensure steady
based on the Soyster for optimal road
traffic conditions for
Heuristic. maintenance of the city.
Kraljevica, Croatia.
Optimize the vehicles by
garage based on the
Develop snow and ice combined service
Allocations based on
control operations storm time/fuel consumption
Roadway NRI allowed (Dowds, Novak,
specific routes designed to metric. Evaluate the
the state to optimize Scott, & Sullivan,
maximize the efficiency of competing vehicle
both man hours and 2013)
the service provided in terms allocations based on the
fuel consumption.
of man-hours and fuel. speed with which high
priority roads are
serviced.
An improvement was
added to the optimizing
Develop a model to
Implement a synchronized algorithm that resulted (Salazar-Aguilar,
synchronize arc routing
routing problem for the in significant Langevin, &
for winter maintenance
snow plowing operations improvement to the Laporte, 2012)
operations.
efficiency of the
plowing model.

Final Report 3
 Develop an integrated
Optimize the routes and fleet A list of various
algorithm for Missouri
allocations for Missouri conditions and response
DOT to determine the (Jaung, 2011)
DOT to provide a sufficient as needed for the
most efficient route plans
level of service (LOS). Missouri DOT.
and fleet allocations
Develop a methodology
for deployment of
Determine optimal Use of contract
available crews and
workforce planning and shift employees reduces the
equipment to maintain the (Gupta, 2010)
scheduling for snow and ice total cost to Missouri
most efficient
removal. DOT.
implementation of
resources.
A model was developed
Develop an urban
Develop a data model to that incorporated a
transportation network
represent the transportation multimodal network (Mandloi & Thill,
using standard and
network of an urban area to with restrictive 2010)
customized GIS software
be used for route planning. attributes to represent
tools.
real world scenarios.
Develop a model to take Provides a method to
Determine methods for
into account a variety of produce optimal
producing optimal
road and weather deployment schedules (Fu, Trudel, &
deployment schedules to
conditions to aid in and a framework to Kim, 2009)
conduct winter maintenance
winter maintenance compare future
operations.
operations planning. research.
Enhance a decision support
system for assisting the Assign segments of the The Genetic Algorithm
Maryland State Highway treated road network to with First Fit heuristic
Administration's Office of trucks so that the number reduces the number of (Haghani &
Maintenance staff in of trucks is minimized minimum trucks for Hamedi, 2002)
designing snow emergency and all routes are Calvert County from 14
routes for Calvert County, continuous. to 12 trucks (14%).
MD.
Developed a list of
Determine how to use arc
Analyze arc routing steps to use for (Assad & Golden,
routing methods to
methods and applications. conducting arc routing 1995)
determine routing.
methods.
Using historical data,
Develop a model to predict develop a model for each The model assists in
costs and benefits of winter district to accurately estimating the benefits (Haber & Limage,
maintenance operations in predict costs and benefits to safety, travel time, 1990)
the state of Idaho. of winter maintenance and fuel cost.
operations.

Provide a description of a
Use computer programs Provides a list of steps
computer application to (Bodin & Kursh,
to optimize the routes for to create route
assist in determining the 1979)
street sweeping trucks. optimizing models.
routing of street sweepers.

Final Report 4
The findings listed in Table 2.1 on the previous page support the idea that the optimization of ODOTs
winter maintenance fleet may result in cost savings while maintaining current levels of service (LOS)
within the districts. In addition, the findings from the literature review assisted the research team in
developing a methodology to conduct the route optimization within ODOT Districts 1, 2, and 10.

2.2 Route Optimization Tools

The research team performed the route optimization work using ArcGIS, a geographic information system
(GIS) platform developed by Esri (based in Redlands, California) to produce optimized routes in the form
of GIS-based maps. Since ODOT is already familiar with this program, the results of the proposed project
may be easily incorporated into ODOT’s current maintenance operations. While complex optimization
algorithms are performed in the ArcGIS program, no computations or coding are required by the end
users, which will make it easy for ODOT winter maintenance personnel to implement the optimized
routes.

2.2.1 Network Analyst

Within ArcGIS is the Network Analyst extension which allows users to conduct analyses on
transportation networks (Esri, 2016). Network Analyst was used to create an accurate Network Dataset
that facilitated the optimization of the snow and ice routes within the involved ODOT districts. The
Network Dataset included all roadways within the districts with turning, speed, and elevation data that
accurately represented real world conditions. Further details regarding the creation of the Network
Dataset may be found in Chapter Four of this report.

2.2.2 Vehicle Routing Problem

Upon completion of developing an accurate Network Dataset within the Network Analyst extension, the
Vehicle Routing Problem (VRP) tool was used to generate optimized routes for the three districts
involved with the project. The VRP was initially created by Esri to allow organizations to determine the
most efficient route (or routes when considering a fleet of vehicles) to service orders, thus saving time and
money. For the purposes of this project, the orders that were to be serviced were roadways that ODOT is
responsible to maintain. More details regarding the utilization of the VRP to determine the optimized
routes for snow and winter maintenance may be found in Chapter Four of this report.

Final Report 5
 2.2.3 QTravel
The research team validated the proposed routes from the ROM by utilizing the computer program
QTravel. This program was created by QStarz, a business based in Taipei, Taiwan whose goal is to “bring
GPS and Bluetooth technology into the consumer mainstream” (Qstarz, 2013). The software and GPS
Travel Recorders produced by QStarz were essential in collecting and analyzing the data obtained from
driving the optimized routes. Further details regarding the validation of the optimized routes may be
found in Chapter Four of this report.

Final Report 6
 CHAPTER III PROJECT SETTING

This chapter provides information about the geographical setting for the project. This chapter is divided
into four sections:

 Section One – Project Setting;

 Section Two – ODOT District 1;

 Section Three – ODOT District 2; and

 Section Four – ODOT District 10.

3.1 Project Setting

The route optimization project was conducted in ODOT Districts 1, 2, and 10. As shown in Figure 3.1
below, these districts represent the Northwestern and Southeastern corners of the State, areas that possess
unique geographic and meteorological demands when conducting winter maintenance operations.

Figure 3.1: ODOT Districts involved with the Route Optimization Project.

Final Report 7
Figure 3.2 below shows the average annual snowfall throughout the state of Ohio ( The Ohio Department
of Transportation, 2011). As observed from Figures 3.1 and 3.2, the three districts involved with the
Route Optimization Project do not receive the greatest amounts of snowfall within the state. However, as
described in the following sections of this chapter, the Northwest and Southeast regions of the state
possess unique geographic challenges that may be mitigated through the use and implementation of the
optimized routes derived from this project.

Obtained from ODOT Snow and Ice Practices, 2011

Figure 3.2: Average Yearly Snowfall in Ohio.

The snowfall ranges in these three districts vary from less than 20 inches to 40 inches on average.
However, a wide range of factors impact winter maintenance treatment. More details about these districts
are presented in this chapter.

Final Report 8
 3.2 ODOT District 1

ODOT District 1 is located in the Northwestern Region of Ohio and consists of Allen, Defiance,
Hancock, Hardin, Paulding, Putnam, Van Wert, and Wyandot counties (The Ohio Department of
Transportation, 2016). The geography of the region primarily consists of level terrain as shown in Figure
3.3.

Figure 3.3: District 1 Elevation.

The data showing that the terrain within ODOT District 1 remains relatively level throughout the area
may be found in Figure 3.3. As observed from Figure 3.3, the elevation changes throughout the district
are gradual with the highest elevation of 1,139 ft in the southern area of Harding County to the lowest
elevation of 641 ft in Defiance County. Though this is a 498 ft difference, in comparison to the rest of the
state, and since these changes are gradual, this area is consider level.

The level terrain leaves the district vulnerable to snow blowing onto the roads while conducting snow and
winter maintenance operations. The snow blowing onto the roads presents an operational challenge as
roads must continue to be treated after the snowfall has ceased. An example of snow blowing onto the
roads is shown in Figure 3.4.

Final Report 9
Obtained from Saugeentimes, 2014.

Figure 3.4: Example of Snow Blowing onto Road.

District 1 is responsible for maintaining approximately 3,200 lanes miles of state and federal roadways
(The Ohio Department of Transportation, 2016) as shown below in Figure 3.5.

Figure 3.5: District 1 Routes and Facility Locations.

Final Report 10
In order to effectively maintain these roadways, the district operates with eight county garages and nine
outposts including the future South Wood Outpost and the removal of the Findlay Outpost. Due to the
removal of the Findlay Outpost currently underway, the route optimization within District 1 did not
incorporate the Findlay Outpost in Hancock County and instead implemented the planned South Wood
Outpost. In addition to implementing the South Wood Outpost, the lane addition to I-75 in Hancock
County planned to be constructed in 2017 was also implemented into the Route Optimization project.

3.3 ODOT District 2

ODOT District 2 is located in the Northwestern corner of the state, immediately north of District 1. The
district serves Fulton, Henry, Lucas, Ottawa, Sandusky, Seneca, Williams, and Wood Counties (The Ohio
Department of Transportation, 2016). The geography of the region is similar to District 1 in that the
terrain is relatively level with a difference between the highest and lowest elevations being 606 ft. Figure
3.6 shows the change in elevation throughout the district.

Figure 3.6: District 2 Elevation.

District 2 is responsible for maintaining approximately 3,381 miles of roadways within the district. These
roadways are currently maintained by utilizing eight county garages and two outposts. For the purposes of
this project, the research team optimized the deployment of trucks within the district under three different
scenarios, consisting of implementing new facilities and relocating current county garages in order to
observe any potential time and cost saving by constructing and implementing new garages and outposts.
A summary of the different scenarios are as follows:

Final Report 11
  Part 1 – Current facility locations and additional I-75 lane;

 Part 2 – Additional South Wood Outpost, relocation of Sandusky County Garage, and additional
I-75 lane; and

 Part 3 – Additional South Wood Outpost, relocation of Sandusky and Wood County garage, and
additional I-75 lane addition.

Further details regarding the different scenarios may be found in subsequent subsections of this report.

3.3.1 District 2 Part 1

Part 1 of District 2 consisted of optimizing the routes with an additional I-75 lane and the current garage
locations. This scenario was completed by removing county border restrictions and setting an average
treating speed of 30 mph. The treating speed was determined after surveying county managers on the
current winter maintenance operations, in particular, the typical speeds traveled during snow and ice
events. A district level overview of the facility locations within this scenario is shown in Figure 3.7.

Figure 3.7: District 2 Part 1 Route Optimization Facility Locations.

Figure 3.7 above provides a visual representation of the roadways that must be maintained and the current
facility locations. The data regarding current facility locations and roadways to be treated helped the
research team to ensure an accurate and thorough analysis was conducted for District 2 Part 1.

Final Report 12
 3.3.2 District 2 Part 2
Optimizing the routes within District 2 Part 2 consisted of keeping the additional I-75 lane increase as
observed in Part 1 , a newly constructed outpost in southern Wood County, and a new county garage
location in Sandusky County. The outpost in southern Wood County will be shared amongst Wood
County in District 2 and Hancock County in District 1. The intent of adding this outpost is to ensure that
an adequate LOS is maintained on I-75 in both Districts 1 and 2. Figure 3.8 below provides a district level
overview of the facility locations.

Figure 3.8: District 2 Part 2 Route Optimization Facility Locations.

In order to clearly show the changes that occurred from Part 1 to Part 2, Figure 3.9 and Figure 3.10 show
a zoomed-in map of the changed areas in Wood and Sandusky Counties. Figure 3.9 on the following page
shows the location of the proposed outpost in Wood County.

Final Report 13
Figure 3.9: Additional Outpost in Wood County.

As may be observed from Figure 3.9, the additional outpost to be added to District 2 Part 2 was located at
the northwest corner of Mercer Rd. and Middleton Pike. The outpost was incorporated into the District 2
Part 2 analysis for the potentially increased efficiency in winter maintenance operations. Another aspect
of the District 2 Part 2 route optimization consisted of relocating the Sandusky County Garage, shown in
Figure 3.10.

Final Report 14
Figure 3.10: Change in Sandusky County Garage Location.

Figure 3.10 above shows the current location of the Sandusky County Garage (left) at the northeast corner
of Oak Harbor Road and Sugar Street and the new location north of US-20 along SR 53 (right).

3.3.3 District 2 Part 3

District 2 Part 3 consisted of keeping the additional lane on I-75, the outpost in southern Wood County,
and the new Sandusky County Garage location as seen in Part 2. Part 3 also consisted of moving the
Wood County Garage from its current location to the proposed location on SR 582. Figure 3.11 provides
a district level overview of the facility locations within this scenario.

Final Report 15
Figure 3.11: District 2 Part 3 Route Optimization Facility Locations.

The largest consideration in the Part 3 route optimization is the analysis of the relocation of the Wood
County Garage. In order to better show the relocation of the Wood County Garage, Figure 3.12 provides a
zoomed-in view of the current location and the proposed location of the Wood County Garage.

Figure 3.12: District 2 Part 3 Wood County Garage Change in Location.

Final Report 16
The current location of the Wood County Garage that was utilized for the Parts 1 and 2 analysis is shown
in the zoomed-in section to the left in Figure 3.12. The zoomed-in section to the right in Figure 3.12shows
the proposed location on SR 582.

3.4 ODOT District 10

ODOT District 10 is located in the Southeastern corner of the state and consists of Athens, Gallia,
Hocking, Meigs, Monroe, Morgan, Noble, Vinton, and Washington Counties (The Ohio Department of
Transportation, 2016). As shown in Figure 3.13, the region consists of mountainous terrain with an
elevation difference of 933ft between the highest and lowest points. In addition, Figure 3.2 on page 8
shows that District 10 annually receives less than twenty to thirty inches of snow ( The Ohio Department
of Transportation, 2011). The combination of mountainous terrain and consistent snowfall presents a
challenge for winter maintenance operations in the district due to the winding rural roads, making them
difficult to treat in a safe and timely manner.

Figure 3.13: District 10 Elevation.

Final Report 17
District 10 is responsible for maintaining over 4,000 lane miles of state highways within the district (The
Ohio Department of Transportation, 2016). A district level map of the roadways that the district must
maintain and the facility locations is shown in Figure 3.14.

Note: Some of the refill facilities are located outside of the District but are utilized by District 10.

Figure 3.14: District 10 Routes and Facility Locations.

As may be seen in Figure 3.14, the transportation system within the district consists of winding roads
reflective of the mountainous terrain of the region. In order to ensure the roads in the district are treated
within the acceptable LOS, the district utilizes nine county garages, eight outposts, and three refill
facilities located along the perimeter of the district.

Final Report 18
 CHAPTER IV ROUTE OPTIMIZATION METHODOLOGY

The methodology section is divided into two sections. The first section discusses the general development
of the ROM in ArcGIS, including the input and data requirements. The second section describes the status
of the data collection and inputting for the ROM for this project, including the digitization of ODOT
maintenance routes and information regarding ODOT facilities and the trucks used for winter
maintenance.

4.1 Initial Model Creation

In order to develop the initial ROM, the research team followed the process described in Figure 4.1
below. It is a summary of the process to develop the initial ROM. The following subsections of this report
describe the data collected and the parameters implemented to ensure the ROM’s capability of effectively
optimizing the operational trucks in each district.

Acquire Roads
Dataset

Data Collection

Route Optimization
Model Development

Plowing Route
Cycle Time
Locations Restrictions

Figure 4.1: Initial Route Optimization Model Development.

Final Report 19
The development of the route optimization began with the preparation of data in the form of layers for the
map of the state of Ohio with additional layers for each category of input data. The first additional layer
was the roads dataset for the entire state of Ohio. Once all the roads in Ohio were uploaded, a network
analyst dataset was created in order to define the roadway edges and capture elevation differences. The
importance of defining roadway edges was seen in locations that include a highway overpass: in two
dimensions, a highway overpass appeared to be an intersection with the road beneath it; when the edges
are used by the model, an elevation difference is applied to the two roads, allowing the program to
recognize the road configuration as an overpass rather than an intersection. An example of the roads
dataset is presented in Figure 4.2.

Obtained from ODOT Transportation Information Mapping System (TIMS)

Figure 4.2: Example of Elevation Differences and Directionality for Roads.

In addition to elevation differences, network attributes in the road layer included road hierarchy (freeway,
arterial, collector, or local road), direction of travel (two-way vs. one-way roads), cost attributes (such as
travel time), and distance. The one-way road attribute is utilized to ensure the model does not route a
vehicle in the wrong direction. The directional basis of the roads, which is an especially important
consideration when routing on divided highways, was built into the road layer. Once the network
attributes were defined and the lengths of all roads determined, travel times were calculated based on the
length and the typical speed traveled during snow events for each road segment.

Final Report 20
4.2 Data Collection

The ROM required information on all of the snow and ice routes that ODOT maintains in the state of
Ohio. Districts 1, 2, and 10 provided the route information in a variety of formats, including digital maps
that were created by using spatial software as well as printed maps with hand drawn routes. Once
acquired, the routes were digitized for use as a base map that included all the routes to be optimized. The
digitized ODOT snow and ice routes for the districts involved with the project are shown in Figure 4.3.
This map shows the current routes that each district is responsible to maintain. The format of this map
allowed it to be utilized as a base layer during route optimization. This layer was broken down by district,
and it did not contain any additional information.

Figure 4.3: ODOT Snow and Ice Routes.

Once the routes were digitized, the next step was to locate the garages and outposts within ArcGIS to
begin the optimization process. This information was obtained from ODOT leadership and implemented
into the ROM. The current number of trucks stationed at each garage, outpost, and refill facilities were
inputted to be used in the model. Plowing locations were then added along the snow and ice routes
provided by ODOT so that trucks could be routed from each garage along the routes. For a particular
plowing location, the optimization model was capable of accounting for salt application. Additionally,
each truck was assigned an associated capacity so that the model could account for the fact that a truck
will run out of material and will need to refill its hopper.

Final Report 21
4.3 Creating the Route Optimization Model

After collecting the necessary data to develop the ROM, the research team utilized the available tools
within Esri’s VRP to create and finalize the model before optimizing the routes within the districts. The
following applications were addressed to create and finalize the ROM within each district:

 Loading plowing locations to be used in the VRP;

 Limiting trucks from traveling on county and township roads; and

 Assigning cycle times to ensure LOS was maintained.

4.3.1 Loading Plowing Locations

The data collected on the plowing locations were used to create the ROM in each district. In order for the
VRP to recognize these plowing locations, the research team identified necessary additional steps to
create the ROM. Figure 4.4 below provides an example of the snow and ice routes in District 10 based on
the data collected for this project.

Final Report 22
Figure 4.4: Snow and Ice Routes within District 10.

Even though Figure 4.4 above shows the snow and ice routes within District 10, the VRP tool that
optimized the routes was not able to operate solely on the data shown. In order for the VRP to be able to
use the data, the research team manually inserted point locations along the roadways that the VRP refers
to as “orders”. Figure 4.5 shows the plowing locations of the district as a whole.

Final Report 23
Figure 4.5: Plowing Locations within District 10

As may be observed from Figure 4.5 above, the plowing locations were placed along all roadways that
must be treated within the district during snow and ice events. Within each of these orders were attributes
that allowed the research team to input the application rate of salt to be used within the model as well as
the time windows to prioritize the routes. Both of these parameters were able to be modified to account
for various application rates and the unique cycle time requirements within each district.

4.3.2 Route Restrictions

After uploading the plowing locations and determining the application rate and cycle times, the ROM
ought to be capable of optimizing the routes within the district. However, the model would have allowed
for trucks to travel along all roadways within the district, an action that was determined to be undesirable
after numerous meetings with ODOT leadership. To prevent the trucks within the ROM from traveling
along these county and township roads, the research team implemented route restrictions in each of the

Final Report 24
districts. An example of the route restrictions being implemented in District 10 may be found in Figure
4.6.

Figure 4.6: Example of Route Restriction Implementation in District 10.

The route restrictions shown in Figure 4.6 were polygon barriers that were manually drawn within the
VRP. Once the restrictions were drawn, the VRP was able to optimize the routes within each district
while limiting them from traveling on county and township roads.

4.3.3 Assigning Cycle Times

In general, cycle times for each route within a district were determined from the LOS requirements that
the district must uphold. Because cycle times are directly related to the LOS requirements, the cycle times
were important parameters within the Route Optimization project and thus required diligent inputs into
the ROM. In order to incorporate the cycle times into the model, the research team added attributes to
each plowing location and truck to be used within the model for each district.

Final Report 25
As previously mentioned in this report, each plowing location was assigned a time window in which it
would be able to be treated. This time window pertained to the LOS requirement for the route priority of
each plowing location, which means that priority one plowing locations had a shorter time window than
other plowing locations. That is, priority one routes would be treated in a lesser cycle time than priority
three routes.

Table 4.2: Level of Service within ODOT Districts 1, 2, and 10.

Level of Service
Priority 1 Priority 2 Priority 3
District
(min) (min) (min)
1 60 90 120
2 60 90 120
10 90 120 150

Note: The LOS requirements shown above were

acquired from each District's leadership.
As may be seen from Table 4.2 above, District 1 maintains a LOS requirement that priority one routes
must be treated within 60 minutes while priority three routes were to be treated within 120 minutes. In
order to implement these LOS requirements into the plowing locations within the ROM, the time window
for priority one plowing locations was set to 60 minutes and the priority three time window was set to 120
minutes.

Apart from adding time windows at each plowing location, the research team also limited the time during
which that each truck was available to travel after leaving the facility. Similar to the time windows used
for the plowing locations, the time allotted for each truck was dependent on the priority of the route that
the truck was to treat. Using the same LOS requirements previously described for the plowing locations,
trucks that were to treat priority one routes were allowed to leave the facility for 60 minutes while trucks
for priority three routes were allowed to leave the facility for 120 minutes. The limitation of how long
trucks may be traveling outside of the garage helped to ensure that the LOS was maintained for all
priority routes within each district.

4.4 Initial Route Optimization

In order to optimize the routes within a district, the research team first utilized the ROM with the entire
truck inventory that each district maintained. By first utilizing all trucks within a district, the ROM was
able to provide a baseline of the area that each facility was most capable of maintaining while satisfying
the LOS requirements within the district. The initial route optimization was accomplished by restricting
the trucks from traveling on county and township roads, by normalizing driving at typical treating speeds

Final Report 26
for each district, and by utilizing an application rate of 250 lbs/ln mile. This initial optimization of routes
also took into account the removal of county border limits within the district, thus allowing the ROM to
determine the most efficient routes and treating areas for each facility. By following the parameters
previously described, the research team was able to produce district overview maps, individual route
maps, and individual route descriptions for Districts 1, 2, and 10. Further details regarding the initial
route optimization may be found in Chapter 6 of this report.

4.5 Verification Process

Due to unique challenges each district faces, a specific verification plan was developed for each district to
minimize work disruption and for providing data in a timely manner. All plans involved the initial
optimized routes to be driven at typical treating speeds and an additional iteration of a current route for
snow and winter maintenance. The data for the driven routes were collected from GPS transponders
(QStarz model Travel Recorder XT data loggers, as shown in Figure 4.7) and further analyzed with the
Qtravel software. Further details regarding the verification process may be found in Chapter 7 of this
report.

Figure 4.7: The GPS Transponder for Collecting Data from Driving the Proposed Routes.

Final Report 27
 4.6 Fleet Optimization Methodology

The fleet optimization portion of the Route Optimization project consisted of optimizing the fleet within
each district to determine which garages and outposts may remove trucks and which facilities may need
additional trucks to maintain the current LOS within each district. By determining which facilities may
remove trucks or require additional trucks, ODOT may experience significant cost savings while
continuing to effectively conduct winter maintenance operations. In order to conduct fleet optimization
the research team began with the initial ROM that utilized all of the trucks available within the district
and followed the process described in Chapter 8 of this report to determine the minimum number of
trucks needed to treat all roads within the district under a worst-case scenario of winter maintenance
operations. The parameters used to construct a worst-case scenario of winter maintenance operations
consisted of the desired cycle times (relating to the LOS requirements), the typical driving speeds during
snow events, and the application rate. The desired cycle times and typical driving speeds for winter
maintenance operations are unique for each district, but all districts were optimized to account for a 400
lbs/ln mile application rate. The 400 lbs/ln mile application rate was determined through meetings with
ODOT leadership to be an acceptable application rate to into account a worst-case scenario for winter
maintenance operations. Additional information regarding the Fleet Optimization may be found in
Chapter 8 of this report.

Final Report 28
 CHAPTER V CURRENT ROUTE ANALYSIS

In order to ensure that a thorough analysis was conducted regarding potential time savings from current
winter maintenance operations to the optimized operations, the research team determined that an analysis
of the current routes being utilized from Districts 1,2, and 10 would be necessary. In order to ensure the
current route analysis was properly conducted, the research team followed the process shown in Figure
5.1 below.

Obtain Current
Routes

Digitize Current
Operational Trucks

Implement Current
Operational Trucks
into ROM

Analysis of Current
Operational Trucks

Note: The analysis conducted on the current operational trucks in each district was the same analysis
conducted on the optimized routes discussed in this report.

Figure 5.1: Current Route Analysis Methodology.

The current routes were obtained in two primary formats, 1) maps of routes currently being utilized for
winter maintenance operations and 2) detailed route descriptions that distinguish the start and end points
of each route’s treating area and the designated truck starting facility.

Upon receiving the current routes, the research team digitized the routes in ArcGIS. After completing
digitizing all of the received routes, the research team then generated route descriptions in the same

Final Report 29
format as the optimized routes. The same process to describe the optimized routes was used to calculate
the treating distance, deadhead, and total cycle times of the current routes. Analyzing the current routes in
this manner was conducted for the current routes within district 1, 2, and 10.

This chapter is divided into three sections with each section describing the status of the current routes
within each district. Each section consists of the following sub-sections:

 Amount of operations trucks currently utilized;

 The expected time to treat all roadways once; and

 The percent of routes that satisfy the LOS requirements within the district.

5.1 District 1 Current Routes Overview

Through collaboration with the county managers within District 1, the research team was able to acquire
all routes currently being used to conduct winter maintenance operations within the district. An overview
of the treating areas for each facility within District 1 may be seen in Figure 5.2.

Figure 5.2: District 1 Current Route Overview.

Final Report 30
Figure 5.2 provides a district-level overview of the treating areas for each facility within District 1. The
figure also shows that the routes currently used are primarily restricted to county borders. Despite such
restriction, outposts located along county borders are typically shared amongst the full service facilities
and with treating areas that extend into numerous counties.

5.1.1 District 1 Current Route Analysis

As previously mentioned in this chapter, the current routes were analyzed in the same manner as the
optimized routes presented later in this report. The primary parameters used for this analysis consisted of
the LOS requirements, the average speed a truck travels, the time required for refill, and the capacity of
each truck. In regards to the LOS requirements, the following are the LOS restrictions as determined by
District 1:

 Priority One - 60 minutes;

 Priority Two - 90 minutes; and

 Priority Three - 120 minutes.

The LOS requirements represent the maximum time required for a road to be treated and are a primary
factor in comparing the current routes to the optimized routes. A route was considered to satisfy the LOS
if the truck was able to leave the facility, treat the assigned roadways, return to the facility, and then refill.
If the truck was able to complete a full cycle under the LOS requirements, it was determined that the truck
satisfied the LOS.

In regards to the average speed expected from a truck conducting winter maintenance, District 1
leadership determined that 40 mph for priority one routes and 30 mph for priority two and three routes
were acceptable speeds to use for the route optimization project. These traveling speeds were
implemented into ArcGIS to determine the cycle times for each route.

The capacity of each truck in the district directly relates to the efficiency calculations for the individual
routes, each facility, and the district as a whole. It is important to note that the efficiency is not a fixed
value but rather a range of values dependent on the application rate. The highest efficiency that a truck
may possess incorporates the potential for numerous cycles to be completed at an application rate of 250
lbs/ln mile. The lowest efficiency is due to the worst-case scenario of 400 lbs/ln mile with each truck
being able to complete one cycle before requiring a refill.

From the parameters previously described, the research team was able to determine the results shown in
Table 5.1.

Final Report 31
Table 5.1: District 1 Current Route Analysis.

District 1 Current Route Analysis

Operational Trucks 109
Fleet Size 127
Total Travel Time
7,651
(Minutes)
Percent LOS Maintained 53
District Efficiency Range Low 77
High 87
Note: The operational trucks are trucks that conduct winter
maintenance operations. The fleet size is the total number of
trucks in the district’s truck inventory. The total travel time is the
expected time required to treat all roadways within the district
once. The district efficiency takes into account the worst-case
scenario of treating all routes within the district for one iteration
and allows routes to complete numerous cycles before a truck
returns to the garage for refilling.

As may be observed from Table 5.1 above, District 1 currently utilizes 109 operational trucks with a total
fleet size of 127 trucks. The non-operational trucks are inoperable during winter maintenance due to
mechanical issues. With 109 operational trucks, the district is able to treat all roadways in 7,651 minutes
(127 hours) with 53% of the routes satisfying the district LOS requirements. The district efficiency ranges
from 77% to 86%, depending on the application rate utilized.

5.2 District 2 Current Route Overview

District 2 provided the research team with all routes currently used for winter maintenance operations
except for trucks leaving from Lucas County Garage and Northwood Outpost. Both the Lucas County
Garage and Northwood Outpost experience varied winter conditions that have resulted in constantly
varied routes depending on the conditions of the roadways. Even though routes were not acquired for
these facilities, the treating areas for the remaining facilities were acquired and digitized in ArcGIS. In
addition, the current facility treating areas were obtained as shown in Figure 5.3.

Final Report 32
Figure 5.3: District 2 Current Treating Areas by Facility.

As may be observed from Figure 5.3, the treating areas for each facility are primarily limited to the
county borders within the district.

5.2.1 District 2 Current Route Analysis

The current routes utilized within District 2 were analyzed in the same manner as the optimized routes
presented later in this report. Similar to District 1, the primary parameters used for this analysis consisted
of the LOS requirements, the average speed a truck travels, the time required for refill, and the capacity of
each truck. In regards to the LOS requirements, the following are the LOS restrictions as determined by
District 2:

 Priority One - 60 minutes;

 Priority Two - 90 minutes; and

 Priority Three - 120 minutes.

Final Report 33
In addition to the LOS requirements within the district, the typical traveling speed used throughout the
district during winter maintenance operations was an important parameter when analyzing the current
routes. District 2 leadership determined that 30 mph for all roadways within the district would accurately
represent the average speed traveled while treating the roads. By applying the speed parameter to the
routes received from District 2, the research team was able to calculate the cycle times for each
operational truck within the district.

Table 5.2 shown below provides the number of operational trucks (excluding Lucas County Garage and
North Wood Outpost), fleet size, the expected time to treat all roads once, the percent of routes that
satisfy the LOS requirements, and the district efficiency range.

Table 5.2: District 2 Current Route Analysis.

District 2 Current Route Analysis

Operational Trucks 85
Fleet Size 126

Total Travel Time

7,698
(Minutes)

Percent LOS Maintained 18

Low 76
District Efficiency Range
High 83
Note: The operational trucks in this table do not take into
account trucks from Lucas County Garage or the North
Wood Outpost due to trucks being deployed on an as needed
basis. The operational trucks are trucks that conduct winter
maintenance operations. The fleet size is the total number of
trucks in the district’s truck inventory. The total travel time is
the expected time required to treat all roadways within the
district once. The district efficiency takes into account the
worst-case scenario of treating all routes within the district
for one iteration and allows routes to complete numerous
cycles before a truck returns to the garage for refilling.

Table 5.2 shows that approximately 18% of the routes currently used by District 2 satisfy the district LOS
requirements. This is primarily due to the lack of outposts throughout the district and the LOS
requirements that the district wishes to maintain. By increasing the LOS requirements by thirty minutes
for priority one, two, and three routes, the research team concluded that the percent of trucks that satisfy
the modified LOS requirements within the district increase to 62%.

Final Report 34
In addition to the LOS requirements affecting the percent of routes that satisfy the district LOS
requirements, approximately 50% of the routes obtained from District 2 maintain mixed priority routes.
This affects the percent of routes that satisfy the LOS requirements due to mixed priority routes being
analyzed by the strictest LOS requirements. An example of this is if a route maintains a priority one road
and a priority three road, the LOS requirement for the route relates to priority one.

5.3 District 10 Current Route Overview

District 10 leadership provided all routes currently being used to conduct winter maintenance operations.
The routes were received in a table with specific start points, end points, and starting facility locations.
These data facilitated the digitization of the current routes within District 10 in ArcGIS. The facility
treating areas are shown in Figure 5.4.

Figure 5.4: District 10 Current Treating Areas by Facility.

Similar to Districts 1 and 2, District 10 has primarily restricted the routes to the county borders. The
exceptions are the outposts located near county borders, in which case multiple counties may station

Final Report 35
trucks at the outposts to treat roads in remote areas of the district or provide a higher LOS to priority one
roads.

5.3.1 District 10 Current Route Analysis

Similar to the previous districts, all current routes were analyzed using the same methods as the optimized
routes presented in the following chapters of this report. The parameters include the district LOS
requirements, typical speeds traveled during winter maintenance, the time required for refill, and the
capacity of each truck. The LOS requirements within District 10 differ from those in Districts 1 and 2 due
to the terrain of the area, with the LOS requirements are as follows:

 Priority One - 90 minutes;

 Priority Two - 120 minutes; and

 Priority Three - 150 minutes.

This change in terrain has resulted in a 30-minute increase for each road classification in the LOS
requirement in District 10. The terrain also influences the speed at which trucks travel during winter
maintenance operations with the following being the average speeds for each priority road:

 Priority One – 35 mph;

 Priority Two – 25 mph; and

 Priority Three – 20 mph.

By utilizing the speeds listed above and the data obtained from District 10, the research team was able to
determine the cycle times for each route within the district. The cycle times were then used to determine
the percent of routes that satisfy the district LOS requirements. Table 5.3 provides a summary of the
current trucks used for winter maintenance operations within District 10.

Final Report 36
Table 5.3: District 10 Current Route Analysis.

District 10 Current Route Analysis

Operational Trucks 116
Fleet Size 128

Total Travel Time

12,650
(Minutes)

Percent LOS Maintained 60

Low 81
District Efficiency Range
High 86
Note: The operational trucks are trucks that conduct winter
maintenance operations. The fleet size is the total number of
trucks in the district’s truck inventory. The total travel time is
the expected time required to treat all roadways within the
district once. The district efficiency takes into account the
worst-case scenario of treating all routes within the district
for one iteration and allows routes to complete numerous
cycles before a truck returns to the garage for refilling.

As may be observed from Table 5.3 above, District 10 currently utilizes 116 operational trucks to conduct
winter maintenance operations. The fleet consists of 128 trucks, including those unavailable for use due to
mechanical or other issues. The total time required to treat the district for one iteration is 12,650 minutes
and the percent of trucks that satisfy the LOS throughout the district is 60%. The range of efficiency is
81% at an application rate of 400 lbs/ln mile and 86% at an application rate of 250 lbs/ln mile. The higher
efficiency is observed at a lower application rate due to more cycles being completed before requiring a
refill at the nearest facility.

5.4 Current Route Analysis Summary

In summary, the results presented in this chapter provide the research team with a baseline District LOS
satisfaction requirement that facilitated the optimization of each district’s fleet as discussed in Chapter 8
of this report. Specifically, the current route analysis allowed the research team to determine how many
trucks could be removed from each district’s fleet to maintain a similar LOS. The current route analysis
also allowed the research team to determine the potential time savings to treat each district, a valuable
tool when considering the potential cost savings within ODOT and increases in safety for the drivers
during snow and ice events.

Final Report 37
 CHAPTER VI INITIAL ROUTE OPTIMIZATION

The initial route optimization consisted of utilizing the ROM within the entire truck inventory that each
district maintained. This optimization of the fleet size instead of the number of operational trucks
currently used provided a baseline of the treating areas of each facility as well as the LOS maintained
with the additional trucks utilized to conduct winter maintenance operations. As previously mentioned in
Chapter 4 of this report, the routes were optimized by removing the county border limits, trucks traveling
at typical treating speeds, and utilizing an application rate of 250 lbs/ln mile. By implementing these
parameters into the ROM, the research team produced district overview maps, individual route maps, and
individual route descriptions for Districts 1, 2, and 10. A summary of the initial route optimization
process is shown in Figure 6.1.

Implement Total
Fleet

Truck Decision
Matrix

Optimize Truck
Placement

Implement Safe
Turnaround
Locations

Export Map and

Route Description

Figure 6.1: Initial Route Optimization Methodology.

Final Report 38
An example of the district overview map for District 2 Part 1 is shown below in Figure 6.2.

Figure 6.2: Example of District 2 Part 1 Overview Map.

As shown in Figure 6.2, the district overview map provides the audience with a simple visual aid to see
the area that the ROM initially determined each facility should maintain when utilizing all trucks within
the district. It is important to note that the thicker lines in Figure 6.2 represent multi-lane roads, further
showing where each facility is conducting maintenance operations. District overview maps were created
for all districts involved in the Route Optimization Project.

In addition to determining the optimal treating areas for each facility, the research team provided detailed
maps and route descriptions for the individual routes within each district. Numerous meetings were held
at the district and county levels to take into account areas of unique concern, such as steep inclines
requiring slow speeds or preventing trucks from turning around at potentially dangerous locations. The
comments from these meetings and the implementation of the ROM generated the optimized routes maps
for each route and facility within Districts 1, 2, and 10. An example of the initial optimized route map
that utilized all trucks within the district is shown in Figure 6.3.

Final Report 39
Figure 6.3: Example of a Map produced from the Initial Optimized Routes in Fulton County.

Figure 6.3 shows an example of the facility maps of the individual routes within the district. The maps
were created for all facilities in the districts involved with this project and provide the audience with the
initial description of the optimized routes. In order to provide further details of the routes, the research
team developed a route description template to be used in conjunction with the facility maps. An example
of the route descriptions used to accompany the facility maps is shown in Figure 6.4.

Final Report 40
Figure 6.4: Example of the Route Descriptions used to accompany the Facility Maps.

As shown in Figure 6.4, the route description table provides additional data on each route that is not
available on the facility map. The data included in the route descriptions are the priority of the route, the
total time traveled, deadhead and treated cycle times, the lane miles treated, the efficiency of each route,
and a written description of where each route is treated. The definitions of each of the categories used for
the route descriptions are as follows:

 Route priority – The priority of the roads that the truck maintains;

 Total Time Traveled – The calculated time to complete one cycle;

 Deadhead Time – The time to drive to and from the treating area for each truck; and

 Treated Cycle Time – The time to drive and treat the optimized roads for each truck.

The efficiency of each route is determined from the application rate applied and the capacity of the truck.
This creates a range of efficiencies rather than a specific efficiency value, where the low efficiency is
calculated by determining the amount of cycles that may be completed by using 400lbs/ln mile
application rate and the high efficiency determined from the amount of cycles that may be completed at a
250 lbs/ln mile application rate. The equation used to calculate the low and high efficiencies for each
route is shown in Equation 6.1.

TCT
𝐼𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙 𝑇𝑟𝑢𝑐𝑘 𝐸𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 = 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 6.1
TCT + DHT

Final Report 41
where,

TCT=Treated cycle time, and

DHT= Deadhead time.

The route descriptions allow the observers to quickly analyze the individual routes and obtain an
estimated cycle time and the lane miles treated. In addition, the written description is particularly helpful
in defining where a route begins and ends.

Another aspect that was considered when presenting the initial optimized routes was the amount of
cycles that could be completed before a refill was needed. This analysis was conducted for each route
within the involved districts and determined the number of cycles that could be completed if the truck
assigned to the route was a single or tandem axle truck. The difference between the two types of trucks is
the capacity each possesses: single-axle trucks have an eight ton capacity while tandem axle trucks have
eleven. The variance in capacity is directly related to the amount of cycles that could be completed at
various application rates. Figure 6.5 provides an example of how these data were presented to accompany
the facility maps and route descriptions previously described.

Figure 6.5: Example of Table and Graphs Produced to Show Number of Cycles Before Refill.

Final Report 42
Figure 6.5 above provides an example of the table produced for each facility showing the number of
cycles that may be completed based on the capacity of the truck and the application rate. The data
presented in Figure 6.5 is valuable to the county managers as they determine which routes the trucks with
larger capacities should treat based on the amount of snowfall in an area (ex. If half of the county is
experiencing heavy snowfall while the other half is not, the managers may assign tandem axle trucks to
the side experiencing heavy snowfall and know the amount of cycles they could complete).

Once the initial optimized routes were created and finalized by ODOT leadership, the research team was
able to proceed with developing and implementing a route verification plan for each district. The routes
produced from the initial optimization would be the routes used to determine the accuracy of the ROM.
The maps of the initial optimized routes from a district and facility overview may be found in Appendix
A of this report.

Final Report 43
 CHAPTER VII ROUTE VERIFICATION

In order to validate the accuracy of the cycle times acquired from the VRP, the research team developed
and implemented a plan for each district to have the initial optimized routes driven with GPS units. The
GPS units collected data on the total time to drive the optimized routes and the average speed driven. A
summary of the route verification process is described in Figure 7.1 below.

Initial Optimized
Routes

Issue GPS units

Operators Drive
Optimized Routes

Receive Data

Analyze Data

Figure 7.1: Route Verification Process.

Upon receiving the data, the research team was able to determine routes that maintained contradictory
cycle times by analyzing the percent difference between the model’s predicted times to the actual times
obtained from driving the proposed route. Equation 7.1 was used to determine the percent difference
between the times obtained from driving the routes and the calculated times from the ROM.

Final Report 44
 (𝐷𝑡 – 𝑀𝑡)
𝑃𝑒𝑟𝑐𝑒𝑛𝑡 𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 = 𝑋 100 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 7.1
𝐷𝑡

where,
Dt =Time acquired from driving the route; and
Mt = Calculated time from the ROM.

After applying Equation 7.1 to determine the percent difference between the time obtained from driving
the route and the calculated time from the ROM, each route was classified into three tiers:
 Tier One – Within 10%;

 Tier Two – Within 10-15%; and

 Tier Three – Greater than 15%.

The routes were classified into three tiers to categorize the cycle times that were acceptable from the
cycle times that may need modification within the ROM to the accurate cycle times of the optimized
routes. Routes classified as Tier One were determined to be acceptable as it allowed the model to produce
realistic times while allowing a 10% variance to account for the many variables of winter maintenance
operations. The routes classified as Tier Two were typically driven at speeds slightly slower or faster than
the speeds utilized within the ROM. Routes classified as Tier Three were typically routes that were driven
incorrectly. Routes classified as Tier Two or Three were deemed to be further analyzed as the percent
difference was too great to predict accurate route completion times. This analysis consisted of information
on whether the driver drove the correct route, whether the driver stopped to yield to traffic flow, and
whether the driver drove significantly faster or more slowly than the speed used in the model. The most
common issues were the route being driven incorrectly and at fast speeds.

To ensure that the ROM produced accurate times for the route, the model was run at the average speed
obtained through the verification. After running the model with the speed obtained from driving the route,
Equation 7.1 was again applied to the data to calculate the percent difference. If the percent difference
between the new model time and the actual time acquired from driving the route could be classified as a
Tier One route, the issue was considered resolved. By calculating the percent difference a second time,
the research team was able to show that the model would have produced accurate cycle times had the
speed within the model been the same as the average speed traveled. This verification may be useful in
future ROMs that allow the user to select unique speeds.

Final Report 45
Once the data were collected, tables and graphs were created to show the relationship between the time
obtained from driving the route and the projected time from the model. The results for each district may
be found in the following subsections of this chapter.

7.1 District 1 Route Verification Overview

Figure 7.2 provides a visual representation of the route verification plan for District 1. The district was
divided into two groups (distinguished as either red or blue) with each group containing six GPS
transponders that were used to acquire data from driving the proposed routes. By dividing the district into
two separate groups, the routes were able to be driven in a timely manner without interfering with the day
to day responsibilities of the garages.

Figure 7.2: Map of Verification Plan for ODOT District 1.

As shown in Figure 7.2 above, the county of Van Wert was not incorporated into the previously
mentioned groups. This was due to the fact that the county already possesses GPS/AVL units in its
vehicles. With the technology available and already installed, the research team was able to acquire all of
the Van Wert County data from driving the routes through the ODOT website that manages the data.

7.2 District 1 Route Verification Results

By following the same process described at the beginning of this chapter regarding the analysis of the
data collected from driving the proposed routes, the research team was able to initially determine the
difference between the time produced from the ROM and the time acquired from driving the routes. Of

Final Report 46
the 115 initial optimized routes, data for 115 routes were collected, constituting 100% of all routes. The
results from the initial data gathered are shown in Figure 7.3.

District 1 Route Verification

Initial Data
200

180

160
y = 1.02x
Verification Time (min)

140 R² = 0.48

120

100

80

60

40

20

0
0 50 100 150
ROM Time (min)

Note: The points above the line of best fit as shown in the green oval represent the cycle times obtained
from routes that were driven slower than the expected speeds as determined by District 1 leadership. In
addition, the points within the red oval represent the routes that were driven at a higher average speed.
These variations in speed may have occurred due to the routes being driven in the spring, causing the
speeds to fluctuate due to a higher level of traffic than experienced during the snow and ice season.

Figure 7.3: District 1 Initial Route Verification Data.

Figure 7.3 above shows the graph of the data collected for the initial optimized routes in District 1. The
line of best fit, set to intercept the origin, has a slope of 1.02. This slope represents a 2% difference
between the data collected and the calculated cycle times. In addition, the line of best fit maintains a
coefficient of determination of 48%, meaning that the cycle times produced from the ROM explains 48%
of the variability between the acquired and calculated cycle times.

The primary reason for the variance in the cycle times acquired from driving the routes and the cycle
times from the ROM is that the operators drove at speeds significantly higher than the speeds utilized in
the ROM due to safety concerns regarding driving heavily traveled roads in the spring and summer
seasons. To determine if the ROM was capable of determining accurate cycle times from the speed
utilized within the model, the research team ran the ROM with the average speed acquired from driving

Final Report 47
the initial optimized routes. The results of reconfiguring the ROM to utilize the higher speeds are shown
below in Figure 7.4.

District 1 Route Verification

Reconfigured Data
200

180

160
Verification Time (min)

140 y = 1.01x
R² = 0.99
120

100

80

60

40

20

0
0 50 100 150
ROM Time (min)

Figure 7.4: District 1 Reconfigured Verification Data.

Adopting the higher speeds in the ROM resulted in a 1% difference between the cycle times calculated
from the ROM and the time acquired from driving the initial optimized routes. In addition, the coefficient
of determination increased from the previous 62% to 99%, that is, 99% of the variance is now predictable
when comparing the ROM cycle times to the acquired cycle times. The results of this analysis show that
the speeds utilized in the ROM will accurately calculate the real world cycle times if the speeds used in
the ROM are the same as those used for real world snow and ice events.

7.3 District 2 Route Verification Overview

Similar to District 1, District 2 was divided into separate groups in order to facilitate the route verification
data collection (as shown in Figure 7.5). Of the three groups, the first consisted of Williams, Fulton, and
Henry Counties. The second group consisted of Lucas and Wood Counties; and the third group of Ottawa,
Sandusky, and Seneca Counties. The groups were determined by the number of routes that needed to be
driven and the geographic location of the counties in each group.

Final Report 48
Figure 7.5: Map of Verification Plan for ODOT District 2.

7.4 District 2 Route Verification Results

In order to present the route verification data in an organized manner, this section is organized into three
subsections. Each subsection presents the results from each scenario analyzed for District 2 and contains
information regarding the initial data obtained and the reconfigured cycle times.

7.4.1 District 2 Part 1 Route Verification Results

The results from the route verification process for District 2 Part 1 showed that 126 of the 126 routes were
driven. The route verification results from District 2 Part 1 have resulted in 100% of the routes were
classified as Tier One after operating the ROM at the speeds acquired from driving the routes. Figure 7.6
shows the data acquired before running the model at the average speeds driven.

Final Report 49
 District 2 Part 1 Route Verification
Initial Data
200
y = 1.02x
180 R² = 0.33
160
Verification Time (min)

140

120

100

80

60

40

20

0
0 50 100 150
ROM Time (min)

Figure 7.6: District 2 Part 1 Initial Route Verification Data.

As may be observed from Figure 7.6, the initial data obtained driving the routes does not match the line of
best fit, as shown by the coefficient of determination being 33%. The primary reason for the variance in
the cycle times acquired from driving the routes and the cycle times from the ROM is that the operators
drove at speeds significantly higher than the speeds utilized in the ROM due to safety concerns of driving
at winter maintenance speeds during the summer months. In addition, there was a significant amount of
road construction projects being completed in District 2 during the time of verifying the routes, resulting
in sporadic cycle times. To determine if the ROM was capable of determining accurate cycle times from
the speed utilized within the model, the research team ran the ROM with the average speed acquired from
driving the initial optimized routes. The results of reconfiguring the ROM to utilize the higher speeds are
shown in Figure 7.7.

Final Report 50
 District 2 Part 1 Route Verification
Reconfigured Data
200

180

160
y = 1.03x
Verification Time (min)

140 R² = 0.98

120

100

80

60

40

20

0
0 50 100 150
ROM Time (min)

Figure 7.7: District 2 Part 1 Reconfigured Verification Data.

Adopting the speeds obtained from driving the routes in the ROM resulted in a 3% difference between the
cycle times calculated from the ROM and the time acquired from driving the initial optimized routes. In
addition, the coefficient of determination increased from the previous 33% to 98%, meaning 98% of the
variance is now predictable when comparing the ROM cycle times to the acquired cycle times. The
results of this analysis show that the speeds utilized in the ROM will accurately calculate the real world
cycle times if the speeds used in the ROM are the same as those used for real world snow and ice events.

7.4.2 District 2 Part 2 Route Verification Results

The results from the route verification process for District 2 Part 2 showed that 126 of the 126 routes were
driven. The route verification results from District 2 Part 2 have resulted in 100% of the routes were
classified as Tier One after operating the ROM at the speeds acquired from driving the routes. Figure 7.8
shows the data acquired before running the model at the average speeds driven.

Final Report 51
 District 2 Part 2 Route Verification
Initial Data
200

180 y = 1.01x
R² = 0.33
160
Verification Time (min)

140

120

100

80

60

40

20

0
0 50 100 150
ROM Time (min)

Figure 7.8: District 2 Part 2 Initial Route Verification Data.

Similar to the route verification results presented in District 2 Part 1, the initial data obtained from driving
the routes does not match the line of best fit, as shown by the coefficient of determination being 33%. The
variance in the cycle times acquired from driving the routes and the cycle times from the ROM is that the
operators drove at speeds significantly higher than the speeds utilized in the ROM due to safety concerns
of driving at winter maintenance speeds during the summer months. The road construction projects being
completed in District 2 also resulted in higher cycle times. To determine if the ROM was capable of
determining accurate cycle times from the speed utilized within the model, the research team ran the
ROM with the average speed acquired from driving the initial optimized routes. The results of
reconfiguring the ROM to utilize the acquired speeds are shown below in Figure 7.9.

Final Report 52
 District 2 Part 2 Route Verification
Reconfigured Data
200

180

160
y = 1.02x
Verification Time (min)

140 R² = 0.98

120

100

80

60

40

20

0
0 50 100 150
ROM Time (min)

Figure 7.9: District 2 Part 2 Reconfigured Verification Data.

Implementing the speeds obtained from driving the routes into the ROM resulted in a 2% difference
between the cycle times calculated from the ROM and the time acquired from driving the initial
optimized routes. In addition, the coefficient of determination increased from the previous 33% to 98%,
that is, 98% of the variance is now predictable when comparing the ROM cycle times to the acquired
cycle times. The results of this analysis show that the speeds utilized in the ROM will accurately calculate
the real world cycle times if the speeds used in the ROM are the same as those used for real world snow
and ice events.

7.4.3 District 2 Part 3 Route Verification Results

The results from the route verification process for District 2 Part 3 showed that 126 of the 126 routes were
driven. The route verification results from District 2 Part 3 have resulted in 100% of the routes were
classified as Tier One after operating the ROM at the speeds acquired from driving the routes. Figure 7.10
shows the data acquired before running the model at the average speeds driven.

Final Report 53
 District 2 Part 3 Route Verification
Initial Data
200

180 y = 1.01x
R² = 0.29
160
Verification Time (min)

140

120

100

80

60

40

20

0
0.0 50.0 100.0 150.0
ROM Time (min)

Figure 7.10: District 2 Part 3 Initial Route Verification Data.

Similar to the route verification results presented in District 2 Parts 1 and 2, the initial data obtained from
driving the routes does not match the line of best fit, as shown by the coefficient of determination being
29%. The variance in the cycle times acquired from driving the routes and the cycle times from the ROM
is that the operators drove at speeds significantly higher than the speeds utilized in the ROM due to safety
concerns of driving at winter maintenance speeds during the summer months. The road construction
projects being completed in District 2 also resulted in higher cycle times. To determine if the ROM was
capable of determining accurate cycle times from the speed utilized within the model, the research team
ran the ROM with the average speed acquired from driving the initial optimized routes. The results of
reconfiguring the ROM to utilize the acquired speeds are shown in Figure 7.11.

Final Report 54
 District 2 Part 3 Route Verification
Reconfigured Data
200
y = 1.03x
180 R² = 0.98

160
Verification Time (min)

140

120

100

80

60

40

20

0
0 50 100 150
ROM Time (min)

Figure 7.11: District 2 Part 3 Reconfigured Verification Data.

Implementing the speeds obtained from driving the routes into the ROM resulted in a 3% difference
between the cycle times calculated from the ROM and the time acquired from driving the initial
optimized routes. In addition, the coefficient of determination increased from the previous 29% to 98%,
that is, 98% of the variance is now predictable when comparing the ROM cycle times to the acquired
cycle times. The results of this analysis show that the speeds utilized in the ROM will accurately calculate
the real world cycle times if the speeds used in the ROM are the same as those used for real world snow
and ice events.

7.5 District 10 Route Verification Overview

In regards to the route verification plan for ODOT District 10, the research team worked closely with the
District leadership and determined that because of the immense responsibilities of the garages during the
time of the verification, it would be best to collect data for each county individually. The first county to
drive the proposed routes was Gallia County located in the southernmost corner of the district. As each
county finished driving the routes, a member of the research team drove to District 10 to collect the data
and transport the GPS transponders to the next county. The GPS transponders were moved north until

Final Report 55
reaching Monroe County, at which point they were returned to The University of Akron. The movement
of the GPS transponders from county to county is shown in Figure 7.12.

Figure 7.12: Map of Verification Plan for ODOT District 10.

As may be observed from Figure 7.12, Washington County did not receive the GPS units since the trucks
there were already equipped with GPS/AVL systems (similar to Van Wert County in District 1). The
routes in Washington County were not driven due to safety concerns out of respect for the operators who
would have been driving slowly due to the traffic on the roads during the spring and summer months.

7.6 District 10 Route Verification Results

The results from the route verification process for District 10 showed that 90 of the 115 routes were
driven, or approximately 78% of the initial optimized routes. The reason that not all of the routes were
driven was the safety concerns for the operators who would not be able to safely travel the routes at
typical treating speeds during the fall and summer seasons due to increased traffic volumes. Another
reason that explains why 78% of the routes were driven was that the 90 routes that were driven were
classified as Tier One after operating the ROM at the speeds acquired from driving the routes, in which
case both the research team and ODOT leadership decided it was acceptable to validate the ROM-
calculated cycle times. Figure 7.13 shows the data acquired before running the model at the average
speeds driven.

Final Report 56
 District 10 Route Verification
Initial Data
200

180
y = 0.93x
160 R² = 0.55
Verification Time (min)

140

120

100

80

60

40

20

0
0 50 100 150 200
ROM Time (min)

Note: The points above the line of best fit as shown in the green oval represent the cycle times obtained
from routes that were driven slower than the expected speeds as determined by District 10 leadership. In
addition, the points within the red oval represent the routes that were driven at a higher average speed.
These variations in speed may have occurred due to the routes being driven in the spring, causing the
speeds to fluctuate due to a higher level of traffic than experienced during the snow and ice season.

Figure 7.13: District 10 Initial Route Verification Data.

Figure 7.13 above graphically shows the initial data collected that maintain a coefficient of determination
of approximately 55%, which means that the ROM explains 55% of the variability of the data. The slope
of the line shows that there is about a 7% difference between the time acquired from driving the routes
and the calculated time from the ROM. The variance observed from the initial data was due to operators
driving the routes at speeds that were different from the speeds used in the model. In order to analyze the
ability of the ROM to produce accurate cycle times based on the inputted actual speed, the research team
then ran the ROM with the average speeds acquired from driving the routes. The results after
reconfiguring the model to use the acquired speeds are shown in Figure 7.14.

Final Report 57
 District 10 Route Verification
Reconfigured Data
200

180
y = 1.01x
160 R² = 0.97
Verification Time (min)

140

120

100

80

60

40

20

0
0 50 100 150 200
ROM Time (min)

Figure 7.14: District 10 Reconfigured Verification Data.

The reconfigured results show that there is an approximate coefficient of determination value of 97%.
That is, the ROM takes into account 97% of the variance between the cycle times acquired from driving
the proposed routes and the cycle times produced from the model. In addition, the slope of the line shows
that there is approximately a 1% difference between the cycle times acquired and the calculated cycle
times when used the same speeds. The results from verifying the routes supported the concept that the
ROM is capable of producing accurate cycle times when inputting the same speeds as would be traveled
in real world applications.

Final Report 58
 CHAPTER VIII FLEET OPTIMIZATION

The fleet optimization section summarizes the final results obtained regarding the optimized winter
maintenance operations in Districts 1, 2, and 10. As mentioned in Chapter 4 of this report, the
optimization of the fleets consisted of removing the least efficient trucks from the initial optimized routes
in order to determine the minimum number needed at each facility to maintain the current LOS in each of
the districts. This analysis was completed by using the LOS requirements for each district, the typical
treating speeds, the removal of county borders, and utilizing an application rate of 400 lbs/ln mile. The
400 lbs/ln mile takes into account the worst-case scenario regarding the maximum application rate, thus
allowing the ROM to determine an accurate amount of operational trucks needed to best treat the roads
within each district. The flowchart below in Figure 8.1 describes the methodology to optimizing the fleet
within the associated districts.

Optimized Routes for Current

Amount of Trucks

Remove Trucks until Maximum LOS

is Satisfied

Determine Minimum Amount of

Trucks for Each Facility

Create Two Separate Models for Each

Facility

Priority Two and Three Remove Trucks to 40%

It is recommended that trucks Route Model of Minimum Needed
assigned to priority two and
three routes be removed before
removing trucks from the
priority one routes. Priority One Route Remove Trucks to 40%
Model of Minimum Needed

Figure 8.1: Fleet Optimization Methodology.

Final Report 59
Figure 8.1 provides a visual representation of the fleet optimization process that was conducted in all
districts associated with the project. Fleet optimization began by taking the ROM that utilized all trucks in
a district and removing the trucks maintaining the least efficient routes at each facility. After removing
the trucks, the ROM was run and new routes were created to accommodate the missing trucks. The
process of removing trucks assigned to the least efficient routes and running the optimization model was
repeated until the current LOS requirements were satisfied in the district. After hitting the threshold of the
maximum number of trucks that could be removed within the scope of the LOS requirements, the
research team was able to determine the minimum amount of routes needed within the district and each
facility to maintain the desired LOS.

It is important to note that the efficiency for fleet-optimized trucks falls within a range of values and is
not a fixed number. To account for this range, the research team took into account the number of cycles
that may be completed for each truck at an application rate of 400 lbs/ ln mile and 250 lbs/ln mile. The
lower end of the efficiency range was calculated at the number of cycles that may be completed at the
higher application rate and the higher end of the efficiency utilized the lower application rate. Both
efficiencies were calculated for the individual trucks with the equation shown in Equation 6.1 on page 41.
The range of the potential efficiencies for each truck was then used to calculate the overall district
efficiency range. The equation used to calculate the district efficiency is shown below in Equation 8.1;

Σ 𝐼𝑅𝐸
𝐷𝑖𝑠𝑡𝑟𝑖𝑐𝑡 𝐸𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 = 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 8.1
𝑇
where,

IRE=Individual route efficiencies, and

T=Number of operational trucks.

After determining the minimum number of routes for the district, the research team continued to remove
trucks from the ROM to conduct an analysis of the change in cycle times, deadhead times, and lane miles
treated for each route. The research team decided to continue removing trucks from each facility until
reaching approximately 40% of the minimum needed to satisfy the LOS requirements within the district.
The research team believed that the removal of trucks until reaching 40% of the minimum amount needed
would provide enough data to conduct an initial analysis of how the routes change to accommodate the
missing truck and provide a valuable tool for county managers as they determine how to allocate their
limited resources to best conduct winter maintenance operations. It is important to note that trucks
assigned to the priority two and three routes were removed before the trucks assigned to the priority one
routes. This was done to ensure that the priority one routes in the district is effectively maintained,

Final Report 60
regardless of the resources available at each facility at any given time. Figure 8.2 below shows an
example of the fleet optimization process for Hardin County in District 1. After optimizing the district
fleet, the ROM determined that the Hardin County Garage required a seven truck crew as the minimum
number of trucks needed to best satisfy the LOS requirements of the district.

Figure 8.2: Example of the Minimum Truck Crew for Hardin County Garage.

The area that the Hardin County Garage maintains in Figure 8.2 is the same area that must be treated after
trucks are removed from the ROM. This was done to prevent any confusion of where to treat amongst the
different facilities in the district. Once the minimum number of trucks was determined for each facility,
one truck that treated priority two and three routes was removed from the ROM and new routes were
determined to accommodate the removed truck. Figure 8.3 provides an example of how the area treated
by the Hardin County Garage remained the same as well as how the priority two and three routes changed
to accommodate the removed truck.

Final Report 61
Figure 8.3: Example of Hardin County Garage Secondary Routes with One Truck Removed.

When comparing the routes between Figure 8.2 and Figure 8.3, the priority two and three routes changed
to accommodate the removed truck, where Hardin – 07 in Figure 8.2 was absorbed by Hardin – 05 and
Hardin – 06 in Figure 8.3. It may be observed that the priority one trucks (Trucks one through four) from
Figure 8.2 remained unchanged in Figure 8.3. It goes to show that the priority one routes remain the same
and only the priority two and three routes have changed to accommodate the removed truck. The research
team continued to remove trucks from the priority two and three routes until reaching approximately 40%
of the minimum number of trucks needed to satisfy the District’s LOS requirements. This process was
then completed for the priority one routes. The maps produced as a result of removing trucks from both
the priority two and three routes before removing trucks from the priority one routes provide ODOT with
a “playbook” of routes that offers flexible options for the end users to make informed decisions on winter
maintenance operations.

The following sections of this chapter summarize the results for each district after conducting fleet
optimization.

Final Report 62
 8.1 District 1 Fleet Optimization Results

Following the process described in the beginning of this chapter, the fleet optimization results in Table
8.1 below show a comparison in the number of trucks, the district total travel times to treat all roads for
one iteration, the percent of trucks that satisfies the LOS requirements, and the district efficiency ranges
for the trucks currently being driven and the optimized trucks.

Table 8.1: District 1 Route Optimization Summary of Results.

District 1 Fleet Optimization Analysis

Current Fleet Optimized Difference

Operational Trucks 109 104 -5

Fleet Size 127 115 -12

Total Travel Time
7,651 7,443 -208
(Minutes)
Percent LOS
53 53 0
Maintained
District Efficiency Low 77 81 4
Range High 87 88 1
Note: The fleet size takes into account a 10% safety factor due to approximately 10% of the
fleet being unavailable for winter maintenance operations at any given time. The total travel
time is the expected time required to treat all roadways within the district once. The district
efficiency takes into account the worst-case scenario of treating all roads within the district for
one iteration and allowing trucks to complete numerous cycles before returning to the garage
for refilling.
The number of operational trucks shown in Table 8.1 above represents the number of trucks needed to
conduct winter maintenance operations at the same LOS currently being maintained. This number does
not take into account trucks that are unavailable for winter maintenance due to mechanical issues. The
number shown in the Fleet Size row is the recommended fleet size after optimizing the district and adding
the 10% safety factor to account for unavailable trucks at any given time. After optimizing District 1, the
ROM suggests that the district requires 114 trucks to effectively treat all roadways within the district.

The total travel time represents the time required to treat each road within the district once with a 196
minute decrease in comparison with the current winter maintenance operations. The total travel time
includes the deadhead time required to reach the treating area, the time required to treat the determined
area, and the deadhead time to return to the facility from which the truck originated.

Upon applying the same parameters to the current winter operations as used in the optimized operations,
the research team was able to determine that approximately 53% of the trucks currently used satisfy the

Final Report 63
LOS requirements in District 1. After conducting route optimization for District 1, it was found that the
same LOS could be satisfied with a fleet size of 114 trucks.

When calculating the district efficiency, it is important to note that the efficiency falls within a range of
values and is not a fixed number. To account for this range, the research team took into account the
number of cycles that may be completed for each truck at an application rate of 400 lbs/ ln mile and 250
lbs/ln mile. The lower end of the efficiency range was calculated at the number of cycles that may be
completed at the higher application rate and the higher end of the efficiency utilized the lower application
rate. Both efficiencies were calculated for the individual trucks with the equation shown previously in
Equation 6.1 on page 41.

An additional aspect to be considered is the number of trucks that satisfy the LOS requirements within the
district. The LOS requirements change by the priority of the road with priority one roads maintaining a
lower time limit than priority three roads. The LOS requirements in District 1 are as follows:

 Priority One Routes - 60 minutes;

 Priority Two Routes - 90 minutes; and

 Priority Three Routes - 120 minutes.

The cycle times used to determine if the LOS requirements were satisfied for each truck within the district
included the deadhead time to travel to and from the treated area, the time required to treat the determined
section of road, and a refill time of ten minutes. The cycle times for the current, the initially optimized,
and the fleet optimized trucks were calculated and the percent of trucks meeting the LOS requirements
determined. The results of this analysis are shown in Figure 8.4.

Final Report 64
 Number of Operational Trucks vs. Percent LOS Maintained
70

60
Y Maximum LOS
X
Percent LOS is Satisfied

50

40

30

20

10

0
80 85 90 95 100 105 110 115 120
Number of OperationalTrucks

Truck Removal Optimization Current District 1 Operational Trucks

Fleet Optimized Operational Trucks Initial Optimized Operational Trucks

Figure 8.4: District 1 LOS Analysis.

As may be observed in Figure 8.4, District 1 currently operates 109 operational trucks with 53% of the
trucks satisfying the LOS requirements within the district. After conducting Fleet Optimization, the
research team determined that the same LOS may be obtained with 104 operational trucks. Line “X” in
Figure 8.4 above provides a visual explanation of how the number of operational trucks may be decreased
from 109 to 104 while maintaining the same LOS. It is important to note that if additional trucks are
implemented, the maximum LOS that may be satisfied is approximately 65%, as shown by the green
circle labeled “Maximum LOS” in Figure 8.4. This may be observed as the percent of trucks satisfying
the LOS requirements within the district plateaus after additional trucks are added past 113. This plateau
occurs after implementing operational constraints, such as safe turn-around locations and ensuring an
acceptable amount of lane miles is maintained.

A summary of the recommended number amount of operational trucks at each facility may be found in
Table 8.2.

Final Report 65
Table 8.2: Recommended Operational Trucks at Each Facility within District 1.

Fleet Optimized
Garage/Outpost
Operational Truck Count
Allen 6
Beaverdam 5
Delphos 4
Fourth Street 4
Defiance 9
Hicksville 2
Hancock 13
South Wood 4
Hardin 7
Forest 5
Roundhead 2
Paulding 10
Putnam 9
Van Wert 11
Wyandot 9
Carey 4
District Operational
104
Trucks
With 10% Safety 115
Note: The district operational trucks are the sum of
recommended operational trucks at each facility within
the district. Through discussions with ODOT leadership,
it is recommended that the fleet size be determined after
adding an additional 10% to the recommended total of
operational trucks. This additional 10% takes into
account trucks that may be inoperable during snow and
ice events.

It is important to note that the recommended fleet size of 115 trucks takes into account a 10% safety
factor to account for trucks that may be inoperable during snow and ice events due to mechanical or
maintenance issues. A district overview of the District 1 optimized fleet may be found in Appendix B of
this report.

Final Report 66
 8.1.1 District 1 Alternative Fleet Sizes
While conducting fleet optimization within District 1, the research team determined that District 1 may
have options regarding their fleet size depending on its goals and intentions. The first option consists of
maintaining the current number of operational trucks for winter maintenance operations. This is shown
from line “Y” in Figure 8.4. Table 8.3 below summarizes the findings if the current number of operational
trucks were optimized in District 1.

Table 8.3: Level of Service Increase if Current Operational Trucks are Optimized.

District 1 LOS Analysis Under Current Amount of Operational Trucks

Current Optimized Difference
Operational Trucks 109 109 0
Fleet Size 127 120 -7
Total Travel Time
7,651 7,636 -15
(Minutes)
Percent LOS Maintained 53 60 7
District Efficiency Range
Low 77 80 3
District Efficiency Range
High 87 88 1
Note: The optimized fleet size takes into account the number of operational trucks with an
additional 10% safety factor due to approximately 10% of the fleet being unavailable for winter
maintenance operations at any given time. The total travel time is the expected time required to
treat all roadways within the district once. The percent LOS maintained is the percent of trucks
that satisfy the LOS requirements in the district. The district efficiency takes into account the
worst-case scenario of treating all routes within the district for one iteration and allowing
routes to complete numerous cycles before returning to the garage for refilling.

As may be observed from Table 8.3 above, a fleet of 120 trucks with 109 operational trucks increases the
percent of trucks satisfying the District LOS requirements by 7%. This increases the current percent of
trucks satisfying the LOS requirements from 53% to 60%.

The second option would allow District 1 to reach the maximum level of service attainable under the
current facility locations and safe turn-around locations. This analysis summarizes the findings shown in
Figure 8.4 that related the percent of trucks satisfying the LOS requirements in the district to the number
of operational trucks implemented. Table 8.4 summarizes the findings to reach the maximum percent of
trucks satisfying the LOS requirements in District 1.

Final Report 67
Table 8.4: District 1 Maximum Level of Service Attainment.

District 1 Maximum LOS Attainment

Current Optimized Difference
Operational Trucks 109 113 4
Fleet Size 127 125 -2
Total Travel Time
7,651 7,692 41
(Minutes)
Percent LOS Maintained 53 65 12
Low 77 79 2
District Efficiency Range
High 87 88 1
Note: The optimized fleet size takes into account the number of operational trucks with an
additional 10% safety factor due to approximately 10% of the fleet being unavailable for
winter maintenance operations at any given time. The total travel time is the expected time
required to treat all roadways within the district once. The percent LOS maintained is the
percent of trucks that satisfy the LOS requirements within the district. The district efficiency
takes into account the worst-case scenario of treating all routes within the district for one
iteration and allowing routes to complete numerous cycles before returning to the garage for
refilling.

Table 8.4 above shows that a fleet of 125 trucks with 113 operational trucks is needed to reach the
maximum amount of trucks satisfying the LOS requirements. The LOS increases from 53% of the trucks
satisfying the LOS to 65%.

8.1.2 District 1 Fleet Optimization Results Summary

The results previously presented after conducting fleet optimization within District 1 provide an analysis
under three scenarios. A visual representation of the three scenarios is shown below in Figure 8.5.

Fleet Optimization

Current Operational
Current LOS Maximum LOS
Trucks

Figure 8.5: Fleet Optimized Scenarios.

Final Report 68
The first scenario consisting of the optimal fleet size to maintain the LOS that is currently maintained, the
second being the attainable LOS if the current fleet remains unchanged, and the third consisting of the
fleet size to attain the highest LOS within the district. A summary of the assignment of operational trucks
throughout the district is shown in Table 8.5.

Table 8.5: District 1 Operational Truck Assignments for Desired Outcomes.

Current
Garage/Outpost Current LOS Operational Maximum LOS
Trucks
Allen 6 6 6
Beaverdam 5 5 5
Delphos 4 4 5
Fourth Street 4 4 4
Defiance 9 9 9
Hicksville 2 3 3
Hancock 13 13 14
South Wood 4 4 4
Hardin 7 7 7
Forest 5 5 5
Roundhead 2 2 2
Paulding 10 11 12
Putnam 9 10 11
Van Wert 11 12 12
Wyandot 9 10 10
Carey 4 4 4
District
Operational 104 109 113
Trucks
With 10% Safety 115 120 125
Note: The district operational trucks are the sum of recommended operational
trucks at each facility within the district. Through discussions with ODOT
leadership, it is recommended that the fleet size be determined after adding an
additional 10% to the recommended total of operational trucks. This additional
10% takes into account trucks that may be inoperable during snow and ice events.

As may be observed from Table 8.5, many of the facility truck assignments in District 1 remain
unchanged when increasing the LOS. Facilities such as the Putnam and Paulding County Garages
increase, resulting in greater LOS attainment throughout the district.

Final Report 69
 8.2 District 2 Fleet Optimization Results

As previously mentioned in the Chapter 3 of this report, the fleet optimization analysis was conducted
under three difference scenarios within District 2 that utilize potential facility relocations and
construction. The results in this section provide the optimal fleet sizes to maintain a similar LOS that is
currently maintained in the District, the LOS that could be obtained from the implementation of new
facilities with the current fleet, and the fleet size to achieve the maximum LOS attainable in the district
after implementing winter maintenance constraints. This methodology is similar to District 1 as shown in
Figure 8.5 on page 68. The following subsections of this report provide the results for each part in District
2.

8.2.1 District 2 Part 1 Fleet Optimization Results

The fleet optimization results in Table 8.6 show a comparison in the number of trucks, the district total
travel times to treat all roads for one iteration, the percent of trucks that satisfies the LOS requirements,
and the district efficiency ranges for the trucks currently being driven and the optimized trucks.

Table 8.6: District 2 Part 1 Fleet Optimization Results.

District 2 Part 1 Fleet Optimization Analysis

Current
Maximum
Current LOS Operational
LOS
Trucks
Operational Trucks 106 114 123
Fleet Size 117 126 136

Difference From Current

-9 0 10
Fleet Size

Total Travel Time

8,724 8,864 9100
(Minutes)

Percent LOS Maintained 27 38 50

Low 78 77 75
District Efficiency Range
High 87 87 87

Note: The fleet size takes into account a 10% safety factor due to approximately 10%
of the fleet being unavailable for winter maintenance operations at any given time. The
total travel time is the expected time required to treat all roadways within the district
once. The district efficiency takes into account the worst-case scenario of treating all
roads within the district for one iteration and allowing trucks to complete numerous
cycles before returning to the garage for refilling.

Final Report 70
When considering the facility locations in Part 1, the research team has determined that the fleet may be
diminished by 9 trucks to maintain a similar LOS that is currently maintained. In addition, the LOS may
be increased to 38% of the trucks satisfying the District requirements by maintaining the current fleet
size. Lastly, a maximum of approximately 50% of the routes satisfying the LOS requirements may be
observed if increasing the current fleet size by 10 trucks. Figure 8.6 provides a visual representation of
Table 8.6.

Number of Operational Trucks vs. Percent LOS Maintained

60

50
Percent LOS is Satisfied

40

30

Y
20
X
10

0
90 95 100 105 110 115 120 125 130
Number of Operational Trucks

Part 1 Truck Removal Optimization Current District 2 Operational Trucks

Part 1 Maximum LOS Part 1 Fleet Optimized Operational Trucks

Figure 8.6: District 2 Part 1 LOS Analysis.

Figure 8.6 shows the percent of operational trucks that satisfy the district LOS requirements as trucks are
removed. The line denoted with an “X” represents the decrease in the current fleet to maintain a similar
LOS that is currently maintained. The line denoted with a “Y” represents the increase in trucks satisfying
the LOS requirements. The green oval represents the LOS plateau, showing that there is a limit to the
percent of trucks satisfying the LOS requirements after implementing operational constraints.

Table 8.7 has been provided to show the operational truck facility assignment for each desired outcome in
District 2 Part 1.

Final Report 71
Table 8.7: District 2 Part 1 Operational Truck Assignments.

Current
Garage/Outpost Current Maximum
Operational
LOS LOS
Trucks
Fulton 8 9 8
Henry 10 13 13
Lucas 10 10 14
Ottawa 6 6 6
Edison 4 6 6
Sandusky 12 12 14
Seneca 10 10 14
Williams 9 9 11
Wood 20 22 22
Northwood 17 17 15
Southwood 0 0 0
District Operational
106 114 123
Trucks
With 10% Safety 117 126 136
Note: The district operational trucks are the sum of recommended
operational trucks at each facility within the district. Through
discussions with ODOT leadership, it is recommended that the fleet
size be determined after adding an additional 10% to the recommended
total of operational trucks. This additional 10% takes into account
trucks that may be inoperable during snow and ice events.

8.2.2 District 2 Part 2 Fleet Optimization Results

Similar to Part 1, the fleet optimization results in Table 8.8 show a comparison in the number of trucks,
the district total travel times to treat all roads for one iteration, the percent of trucks that satisfies the LOS
requirements, and the district efficiency ranges for the trucks currently being driven and the optimized
trucks under the Part 2 parameters described in Chapter 3 of this report.

Final Report 72
Table 8.8: District 2 Part 2 Fleet Optimization Results.

District 2 Part 2 Fleet Optimization Analysis

Current
Current Maximum
Operational
LOS LOS
Trucks
Operational Trucks 102 114 119
Fleet Size 113 126 131

Difference From
-13 0 5
Current Fleet Size

Total Travel Time

8,422 8,756 8,846
(Minutes)
Percent LOS
28 41 47
Maintained
District Efficiency Low 79 77 75
Range High 88 87 87
Note: The fleet size takes into account a 10% safety factor due to
approximately 10% of the fleet being unavailable for winter maintenance
operations at any given time. The total travel time is the expected time
required to treat all roadways within the district once. The district efficiency
takes into account the worst-case scenario of treating all roads within the
district for one iteration and the best case scenario of allowing trucks to
complete numerous cycles before returning to the garage for refilling.

Table 8.8 above shows that the additional outpost in southern Wood County allow for 13 trucks to be
removed from the current fleet to maintain a similar LOS that is currently maintained. In comparison to
the Part 1 results, the additional outpost allows for an additional four trucks to be removed. This is
primarily due to the outpost’s location allowing for trucks to more efficiently treat priority roads with less
deadhead time. A LOS of 41% may be obtained if the current fleet size is maintained and a LOS of 47%
may be observed if the current fleet is increased by five trucks. A visual representation of the percent of
routes satisfying the LOS requirements based on various operational truck amounts may be observed in
Figure 8.7.

Final Report 73
 Number of Operational Trucks vs. Percent LOS Maintained

60

50
Percent LOS is Satisfied

40

30

Y
20
X
10

0
90 95 100 105 110 115 120 125
Number of Operational Trucks

Part 2 Truck Removal Optimization Current District 2 Operational Trucks

Part 2 Maximum LOS Part 2 Fleet Optimized Operational Trucks

Figure 8.7: District 2 Part 2 LOS Analysis.

Similar to the Part 1 LOS analysis, Figure 8.7 shows the percent of operational trucks that satisfy the
district LOS requirements as trucks are removed. The line denoted with an “X” represents the decrease in
the current fleet to maintain a similar LOS that is currently maintained. The line denoted with a “Y”
represents the increase in trucks satisfying the LOS requirements. The green oval represents the LOS
plateau, showing that there is a limit to the percent of trucks satisfying the LOS requirements after
implementing operational constraints.

Table 8.9 has been provided to show the operational truck facility assignment for each desired outcome in
District 2 Part 2.

Final Report 74
Table 8.9: District 2 Part 2 Operational Truck Assignments.

Current
Current Maximum
Garage/Outpost Operational
LOS LOS
Trucks
Fulton 8 9 8
Henry 10 10 12
Lucas 9 12 12
Ottawa 6 5 6
Edison 4 6 6
Sandusky 12 14 14
Seneca 10 13 14
Williams 9 11 11
Wood 13 13 14
Northwood 15 15 15
Southwood 6 6 7
District Operational
102 114 119
Trucks
With 10% Safety 113 126 131

Note: The district operational trucks are the sum of recommended

operational trucks at each facility within the district. Through
discussions with ODOT leadership, it is recommended that the fleet
size be determined after adding an additional 10% to the recommended
total of operational trucks. This additional 10% takes into account
trucks that may be inoperable during snow and ice events.

8.2.3 District 2 Part 3 Fleet Optimization Results

Similar to the previously described Parts 1 and 2, the fleet optimization results in Table 8.10 show a
comparison in the number of trucks, the district total travel times to treat all roads for one iteration, the
percent of trucks that satisfies the LOS requirements, and the district efficiency ranges for the trucks
currently being driven and the optimized trucks under the Part 3 parameters described in Chapter 3 of this
report.

Final Report 75
Table 8.10: District 2 Part 3 Fleet Optimization Results.

District 2 Part 3 Fleet Optimization Analysis

Current
Current Maximum
Operational
LOS LOS
Trucks
Operational Trucks 101 114 123
Fleet Size 112 126 136

Difference From
-14 0 10
Current Fleet Size

Total Travel Time

8,507 8,617 8,894
(Minutes)
Percent LOS
30 42 51
Maintained
District Efficiency Low 79 77 75
Range High 86 87 87
Note: The fleet size takes into account a 10% safety factor due to
approximately 10% of the fleet being unavailable for winter maintenance
operations at any given time. The total travel time is the expected time
required to treat all roadways within the district once. The district efficiency
takes into account the worst-case scenario of treating all roads within the
district for one iteration and the best case scenario of allowing trucks to
complete numerous cycles before returning to the garage for refilling.

Table 8.10 above shows that the additional outpost in southern Wood County and the relocation of the
Wood County Garage allow for 14 trucks to be removed from the current fleet to maintain a similar LOS
that is currently maintained. In comparison to the Part 2 results, the additional outpost allows for an
additional truck to be removed. This is primarily due to the outpost and county garage’s locations
allowing for trucks to more efficiently treat priority roads with less deadhead time. A LOS of 42% may be
attained if the current fleet size is maintained and a LOS of 51% may be observed if the current fleet is
increased by 10 trucks. It is important to note that the Part 3 results show the largest decrease in fleet size
to maintain the current LOS maintained and also supports the greatest LOS that may be obtained by
increasing the fleet size. A visual representation of the percent of routes satisfying the LOS requirements
based on various operational truck amounts may be observed in Figure 8.8.

Final Report 76
 Number of Operational Trucks vs. Percent LOS Maintained

60

50
Percent LOS is Satisfied

40

30

Y
20
X
10

0
90 95 100 105 110 115 120 125 130
Number of Operational Trucks

Part 3 Truck Removal Optimization Current District 2 Operational Trucks

Part 3 Maximum LOS Part 3 Fleet Optimized Operational Trucks

Figure 8.8: District 2 Part 3 LOS Analysis.

Similar to the LOS analysis for Parts 1 and 2, Figure 8.8 shows the percent of operational trucks that
satisfy the district LOS requirements as trucks are removed. The line denoted with an “X” represents the
decrease in the current fleet to maintain a similar LOS that is currently maintained. The line denoted with
a “Y” represents the increase in trucks satisfying the LOS requirements while maintaining the current
fleet size. The green oval represents the LOS plateau, showing that there is a limit to the percent of trucks
satisfying the LOS requirements after implementing operational constraints.

Table 8.11 has been provided to show the operational truck facility assignment for each desired outcome
in District 2 Part 1.

Final Report 77
Table 8.11: District 2 Part 3 Operational Truck Assignments.

Current
Current Maximum
Garage/Outpost Operational
LOS LOS
Trucks
Fulton 8 9 8
Henry 10 12 13
Lucas 9 10 12
Ottawa 5 5 6
Edison 4 6 6
Sandusky 11 13 14
Seneca 11 12 14
Williams 9 9 11
Wood 12 14 15
Northwood 14 14 14
Southwood 8 10 10
District Operational
101 114 123
Trucks
With 10% Safety 112 126 136
Note: The district operational trucks are the sum of recommended
operational trucks at each facility within the district. Through
discussions with ODOT leadership, it is recommended that the fleet
size be determined after adding an additional 10% to the recommended
total of operational trucks. This additional 10% takes into account
trucks that may be inoperable during snow and ice events.

8.3 District 10 Fleet Optimization Results

By following the same process previously described in this chapter, the research team was able to
determine the optimal fleet size to maintain the current LOS with fewer operational trucks. Table 8.12
shows a comparison in the number of trucks, the district total travel times to treat all routes once, and the
district efficiency ranges for the routes currently being driven and the optimized routes.

Final Report 78
Table 8.12: District 10 Route Optimization Summary of Results.

District 10 Route Optimization Analysis

Current Fleet Optimized Difference
Operational Trucks 116 109 -7
Fleet Size 128 120 -8
Total Travel Time 12,650 11,813 -837
(Minutes)
Percent LOS Maintained 60 61 1
Low 81 83 2
District Efficiency Range
High 86 87 1
Note: The optimized fleet size takes into account the number of operational trucks with
an additional 10% safety factor due to approximately 10% of the fleet being unavailable
for winter maintenance operations at any given time. The total travel time is the expected
time required to treat all roadways within the district once. The percent LOS maintained
is the percent of trucks that satisfy the LOS requirements in the district. The district
efficiency takes into account the worst-case scenario of treating all routes within the
district for one iteration and allowing routes to complete numerous cycles before
returning to the garage for refilling.

As may be observed in Table 8.12, the ROM has determined that the optimal number of operational
trucks in District 10 to be 109. The optimization resulted in a decrease of 7 trucks from the current
number of operational trucks in the district. It is important to note that the number of operational trucks is
not the same as the recommended fleet size in the district since approximately 10% of a district’s fleet
may be inoperable at any given time. Taking the inoperable trucks into account, the research team added
10% to the 109 to show that a truck inventory of 120 trucks is necessary for the district.

The total travel time represents that time required to treat each route once within the district with an 838
minute decrease when comparing the current routes to the optimized routes. The total travel time includes
the deadhead time required to reach the treating area, the time required to treat the determined area, and
the deadhead time to return to the facility where the truck originated. From an operational point of view,
these data are valuable in predicting the amount of time required to treat the entire district but also in
predicting the expected labor costs at a district level.

An additional aspect to be considered is the amount of routes that satisfy the LOS requirements in the
district. The LOS requirements change by the priority of the road with priority one routes maintaining a
lower time limit than priority three roads. The LOS requirements within District 10 are as follows:

Final Report 79
  Priority One Routes - 90 minutes;

 Priority Two Routes - 120 minutes; and

 Priority Three Routes - 150 minutes.

The cycle times used to determine if the LOS requirements were satisfied for each route within the district
included the deadhead time to travel to and from the treated area for each route, the time required to treat
the determined route, and a refill time of 10 minutes. The cycle times for the current routes, the initial
optimized routes, the minimum number of routes, and the fleet optimized routes (removing trucks beyond
the minimum recommended total) were calculated and the percent of routes meeting the LOS
requirements determined. The results of this analysis are shown in Figure 8.9.

Number of Operational Trucks vs. Percent LOS Maintained

70

60
Percent LOS is Satisfied

50 Maximum LOS

40

30

20

10

0
70 75 80 85 90 95 100 105 110 115 120
Number of Operational Trucks

Truck Removal Optimization Current District 10 Operational Trucks

Fleet Optimized Operational Trucks Initial Optimized Operational Trucks

Figure 8.9: District 10 LOS Analysis.

The results shown in Figure 8.9 show that approximately 60% of the routes in the district will satisfy the
LOS requirements when trucks travel at typical winter maintenance operations speeds. The point where
the percentage of the LOS satisfied begins to decrease occurs after the number of routes is diminished to
below 109. In order to increase the percentage of routes satisfying the LOS in the district, facilities should
be constructed in the regions that are currently difficult to maintain. The concept of the district operating
at the maximum LOS attainable is shown in the green oval in Figure 8.9, which highlights the plateau of
the percent of operational trucks that satisfy the LOS in the district. This plateau occurs after

Final Report 80
implementing operational constraints, such as safe turn-around locations and ensuring an acceptable
amount of lane miles is maintained.

A summary of the number of operational trucks per facility is provided in Table 8.13.

Table 8.13: Recommended Operational Trucks at Each Facility in District 10.

Fleet Optimized
Garage/Outpost
Operational Truck Count

Athens 9
Hollister 3
Gallia 11
Hocking 8
Laurelville 2
Meigs 8
Tuppers Plains 5
Monroe 8
Duffy 4
Morgan 10
Noble 14
Vinton 6
Wilkesville 3
Washington 8
Bartlett 2
Belpre 4
Macksburg 4
District Operational
109
Trucks
With 10% Safety 120
Note: The district operational trucks are the sum of
recommended operational trucks at each facility in the
district. Through discussions with ODOT leadership, it is
recommended that the fleet size be determined after
adding an additional 10% to the recommended total of
operational trucks. This additional 10% takes into
account trucks that may be inoperable during snow and
ice events.

Final Report 81
Similar to Districts 1 and 2 previously discussed in this chapter, the recommended fleet size of 120 trucks
takes into account an additional 10% to the recommended number of operational trucks. This addition
ensures there are a sufficient number of trucks to effectively conduct winter maintenance operations as
some trucks may be inoperable during snow and ice events for various reasons, including mechanical
issues. A district overview of the District 10 optimized fleet may be found in Appendix B of this report.

Final Report 82
 CHAPTER IX ROUTE VULNERABILITY

While the Fleet Optimization determined the optimal fleet size to maintain the current LOS within
Districts 1, 2, and 10, the research team was able to identify possible strategies to improve the percent of
routes satisfying the district LOS requirements. These recommendations are the result of recognizing
limitations on the number of operational trucks used to conduct winter maintenance operations from the
available facility locations.

When considering the number of operational trucks satisfying the LOS requirements, there is a pivotal
point when the number of operational trucks will yield the highest LOS satisfaction possible in the
district. This occurs after implementing operational constraints relating to turnaround locations and lane
mile requirements for individual routes. The maximum LOS obtained is limited to the facility locations
within a district due to the remote roadways that may not be treated under the LOS requirements due to
the typical speeds traveled during snow and ice events.

To prove this point, the research team assigned 60 trucks to each facility in Districts 1, 2, and 10 and
utilized the ROM under the average treating speeds each district determined. The results of this analysis
are used to determine the challenging areas in each district where construction of future facilities may
increase the amount of operational trucks that satisfy the LOS.

9.1 District 1 Route Vulnerability

In order to show the areas that would benefit from additional facilities, the research team ran the ROM
that utilized the current facility locations with 60 available trucks at each facility to conduct winter
maintenance operations. By limiting each truck to strictly adhere to District 1 LOS requirements
previously mentioned in this report, the research team was able to determine the challenging areas due to
facility locations as shown in Figure 9.1.

Final Report 83
Figure 9.1: District 1 Demanding Areas.

The areas labeled as “Demanding Areas” were classified as such due to the greatest potential to increase
the number of trucks that would satisfy the LOS requirements if facilities were constructed in these areas.
The challenging areas are difficult to maintain from the current facility locations due to the inability of
trucks to complete the cycles under the LOS requirements.

9.2 District 2 Route Vulnerability

In order to best determine the vulnerability in District 2, this section is divided into three subsections.
Each subsection represents the different scenarios that were analyzed for District 2, relating to the
previously mentioned Parts 1, 2, and 3. Similar to District 1, each scenario in District 2 consisted of
assigning 60 trucks to each facility and allowing the ROM to optimize the routes. The locations not
treated for each scenario were highlighted and presented as an overlay to the district map.

9.2.1 District 2 Part 1 Vulnerability

Upon operating the ROM under the conditions previously described, the research team was able to
identify areas that are difficult to maintain due to the trucks traveling at typical treating speeds during
snow and ice events. Figure 9.2 is a map of the demanding areas that the district currently experiences.

Final Report 84
Figure 9.2: District 2 Part 1 Demanding Areas.

As may be observed from Figure 9.2, there are numerous areas that are difficult to maintain under the
current facility locations. These areas are typically on county borders where no outposts are available to
station trucks, creating deadhead and longer cycle times to treat the roads in the district.

9.2.2 District 2 Part 2 Vulnerability

Under the scenario for District 2 Part 2, which includes the addition of an outpost in southern Wood
County and the relocation of the Sandusky County Garage, it may be observed that that challenging areas
are resolved in the southern area of Wood County and the south eastern area of Henry County. This
observation was determined as a result of comparing the demanding areas from Part 1 to Part 2.

Figure 9.3: District 2 Part 2 Demanding Areas.

Final Report 85
The implementation and construction of the Southwood Outpost resolves some challenging areas but the
majority of the challenging areas in the district remained the same.

9.2.3 District 2 Part 3 Vulnerability

The challenging areas in District 2 under the Part 3 scenario, consisting of moving the Wood County
Garage from its location in Part 2 to a new location in Part 3, may be observed in Figure 9.4. The same
process of deploying 60 trucks at each facility under strict LOS requirements with trucks traveling at
expected speeds during snow events was used to determine these vulnerable areas.

Figure 9.4: District 2 Part 3 Demanding Areas.

Figure 9.4 above shows that the challenging areas along I-75 west of the Northwood Outpost are removed
after implementing the new Wood County Garage location. While this allows for a higher LOS
satisfaction for priority one roads in the district, an additional challenging area arises along US 6 west of
SR 235.

9.3 District 10 Route Vulnerability

As previously mentioned, the research team operated the ROM with 60 trucks in each facility that may be
used to conduct winter maintenance operations. By providing the ROM with excess trucks, the research
team was able to determine the areas that are difficult to maintain due to facility and traveling speed
constraints. The results of this analysis are shown on in Figure 9.5.

Final Report 86
Figure 9.5: District 10 Demanding Areas.

As may be observed from Figure 9.5, there are large demanding areas in District 10. The largest
demanding areas may be found in southern Gallia County, northeastern Washington County, and
northeastern Noble County. Due to the LOS being limited by facility locations rather than the fleet size,
additional facilities in these areas would increase the amount of routes that satisfy the requirements in the
district.

Final Report 87
 CHAPTER X IMPLEMENTATION

This chapter presents the implementation of the optimized routes generated during this study. This
implementation plan was created to assist in successfully implementing the previously discussed results
from this report. This chapter is divided into eight sections:

 Section One – Recommendations for implementation of the optimized routes;

 Section Two – Steps needed to implement the findings from this study;

 Section Three – Suggested time frame for implementation;

 Section Four – Expected benefits from implementation;

 Section Five – Potential risks and obstacles to implementation;

 Section Six – Strategies to overcome potential risks and obstacles;

 Section Seven – Potential users and other organizations that may be affected; and

 Section Eight – Estimated cost of implementation.

10.1 Recommendations for the Implementation of the Optimized Routes

The implementation of the optimized routes is a decision that must first be made at the district level, then
the county level.

10.2 Steps Needed to Implement Findings

The steps needed to implement the findings primarily pertain to the training of personnel, preparations to
station trucks at the optimal facilities, and guidance on future construction of facilities in the demanding
areas outlined in Chapter 9 of this report.

The first step that must be taken to successfully implement the findings of this project is to train county
managers and operators on how to use the optimized routes as a tool to continue providing quality service
during snow and ice events. Care must be taken to show how the optimized routes not only determine the
optimal fleet size in the district involved but also determine the minimum number of operational trucks at
each facility to ensure that an acceptable LOS is maintained.

A concern from many county managers is that they often operate with fewer personnel than called for by
the recommended minimum number of operational trucks. The results of this project have provided
optimized routes that are below the recommended minimum for each facility. This opens the more

Final Report 88
flexibility for county managers to decide how to best allocate their limited resources for conducting snow
and ice operations in their areas of responsibility.

The next step for implementing the findings from this project consists of reallocating trucks to the optimal
facilities as determine through the ROM. A district level plan may be necessary to ensure that all outposts
and county garages have the necessary trucks to adequately treat roads in their areas of responsibility.

The final step consists of using the data presented in Chapter 9 of this report to assist in the future
construction of facilities in each of the district’s challenging areas to increase the LOS maintained.

10.3 Suggested Time Frame for Implementation

The implementation of the optimized routes may begin immediately. The primary time constraint is on
ensuring that the county managers are adequately trained on the utilization of the optimized routes and
those preliminary questions and concerns are addressed prior to the winter season.

10.4 Benefits Expected from Implementation

The expected benefits from the implementation of the findings from this project include cost savings from
a reduced fleet size and a decrease in the amount of time required to treat all roads in Districts 1, 2, and
10. While the question of cost savings is important, it is also worthy of considering the potential increase
in safety for the drivers who travel during snow and ice events. By treating the roads in a shorter amount
of time, the safety for local traffic increases while the risk of accidents related to poor weather conditions
decreases. In addition, the optimization of each facility for a variety of operational trucks available will
enable ODOT to continuing providing outstanding service during snow and ice events, regardless of the
amount of resources available at any given time.

Another benefit that may be gained from the implementation of the findings is the validation of the
challenging areas in each facility. By collecting data from the routes that treat these challenging areas,
ODOT should be able to verify the challenging areas determined by the ROM. The validation of the
challenging areas in the districts involved with the Route Optimization project may justify and guide the
future construction of facilities.

10.5 Potential Risks and Obstacles to Implementation

One of the potential risks of and obstacles to the implementation of the optimized routes is confusion of
the areas each facility is responsible to maintain. The current routes used for winter maintenance have
been used for years and there is a familiarity with the routes that may be difficult to change. This

Final Report 89
familiarity with the current routes may guide operators to revert to the routes they are familiar with and
dismiss the findings of the Route Optimization project.

10.6 Strategies to Overcome Potential Risks and Obstacles

A change in operations is a challenge that all organizations must overcome in their own way, such as
developing a solution that is tailored to the culture and ethos of each organization. The best solution to
overcoming the potential resistance from implementing the findings of this project is to conduct adequate
personnel training before the winter season. This training must include addressing the comments and
concerns regarding the new treating areas for each facility and any questions regarding the treating areas
for the individual routes within the district.

10.7 Potential Users and Other Organizations

The Route Optimization of Districts 1, 2, and 10 may provide information to other organizations that have
access to ArcGIS and conduct winter maintenance operations. These organizations may include
municipalities and state DOTs. In addition, the results of this project may be used to facilitate the
automation of the route optimization process within ArcGIS. By automating the process described in this
report, an organization may experience significant cost savings with minimal manual input upon
successful completion of the automation.

10.8 Estimated Cost of Implementation

ODOT currently has the resources to implement the recommendations from this project at little to no
additional cost.

Final Report 90
 REFERENCES

The Ohio Department of Transportation. (2011). Snow and Ice Practices. The Ohio Department of
Transportation Division of Operations.

Assad, A., & Golden, B. (1995). Chapter 5 - Arc Routing Methods and Applications. In M. B. al,
Handbooks in OR & MS (p. Vol 8).

Bodin, L., & Kursh, S. (1979). A Detailed Description of a Computer System for the Routing and
Scheduling of Street Sweepers. Compur. & Ops Res Vol 6, 131-138.

Dowds, J., Novak, D., Scott, D., & Sullivan, J. (2013). Optimization of Snow Removal in Vermont.
Vermont Agency of Transportation.

Esri. (2016). An Overview of the Network Analyst Toolbox. Retrieved from ArcGIS Pro:
http://pro.arcgis.com/en/pro-app/tool-reference/network-analyst/an-overview-of-the-network-
analyst-toolbox.htm

Fu, L., Trudel, M., & Kim, V. (2009). Optimizing Winter Road Maintenance Operations Under Real-
Time Information. European Journal of Operational Research 196, 332-341.

Gudac, I., Hanak, T., & Marovic, I. (2014). Optimization of Winter Road Maintenance Services Under
Real Time Information. University of Rijeka.

Gupta, D. (2010). Optimal Workforce Planning and Shift Scheduling fo Snow and Ice Removal. Missouri
Department of Transportation.

Haber, D., & Limage, U. (1990). Benefit Cost Analysis of Winter Maintenance Levels of the Idaho
Transportation Department. Boise: Idaho Transportation Department.

Haghani, A., & Hamedi, M. (2002). Calvert County Snow Emergency Decision Support System Phase II.
University of Maryland.

Jaung, W. (2011). Optimizing Winter/Snow Removal Operations in MoDOT St. Louis District - Includes
Outcome Based Evaluation of Operations Final Report. Missouri Department of Transportation.

Mandloi, D., & Thill, J.-C. (2010, March 15). Object-Oriented Data Modeling of an Indoor/Outdoor
Urban Transportation Network and Route Planning Analysis. Geospatial Analysis and Modelling
of Urban Structure and Dynamics, pp. 197-220.

Qstarz. (2013). Retrieved from Qstarz: Beyond Navigation: http://www.qstarz.com/about_Qstarz.html

Salazar-Aguilar, A., Langevin, A., & Laporte, G. (2012). Synchronized Arc Routing for Snow Plowing
Operations. Computer & Operations Research, Volume 39, issue 7.

Final Report 91
The Ohio Department of Transportation. (2016). Department of Transportation District 1. Retrieved from
http://www.dot.state.oh.us/districts/d01/Pages/default.aspx

The Ohio Department of Transportation. (2016). Department of Transportation District 10. Retrieved
from http://www.dot.state.oh.us/districts/d10/Pages/default.aspx

The Ohio Department of Transportation. (2016). Ohio Department of Transportation District 2. Retrieved
from http://www.dot.state.oh.us/districts/d02/Pages/default.aspx

Homer, C.G., Dewitz, J.A., Yang, L., Jin, S., Danielson, P., Xian, G., Coulston, J., Herold,
N.D.,Wickham, J.D., and Megown, K., 2015, Completion of the 2011 National Land Cover
Database for the conterminous United States-Representing a decade of land cover change
information. Photogrammetric Engineering and Remote Sensing, v. 81, no. 5, p. 345-354

Final Report 92
 APPENDIX A: INITIAL OPTIMIZED DISTRICT MAPS

Final Report 93
Final Report 94
Final Report 95
Final Report 96
Final Report 97
 APPENDIX B: FLEET OPTIMIZED DISTRICT MAPS

The individual routes for each facility in District 1 will be provided to the District Highway Maintenance
Administrators in a separate document.

Final Report 98
The individual routes for each facility in District 2 Part 1 will be provided to the District Highway
Maintenance Administrators in a separate document.

Final Report 99
The individual routes for each facility in District 2 Part 2 will be provided to the District Highway
Maintenance Administrators in a separate document.

Final Report 100

The individual routes for each facility in District 2 Part 3 will be provided to the District Highway
Maintenance Administrators in a separate document.

Final Report 101

The individual routes for each facility in District 10 will be provided to the District Highway Maintenance
Administrators in a separate document.

Final Report 102