Scribd
Upload
6 views43 pages

SCM518 2024 Week5 IntegerProgramming (1)

The session focused on integer programming, emphasizing its application in various business scenarios such as capital budgeting, fixed cost manufacturing, and facility location. Key concepts include the necessity for all-or-nothing decisions and the importance of setting up proper constraints to link integer variables with non-integer variables. The session also provided examples and reflections on different models, highlighting the use of integer programming in optimizing decision-making processes.

Uploaded by

4maryjohn4
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
6 views43 pages

SCM518 2024 Week5 IntegerProgramming (1)

The session focused on integer programming, emphasizing its application in various business scenarios such as capital budgeting, fixed cost manufacturing, and facility location. Key concepts include the necessity for all-or-nothing decisions and the importance of setting up proper constraints to link integer variables with non-integer variables. The session also provided examples and reflections on different models, highlighting the use of integer programming in optimizing decision-making processes.

Uploaded by

4maryjohn4
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 43

Session 5

Integer Programming

Yimin Wang
Refresh of last session

 Network models are a special type of LP models


 Graphical representation
 Special solution techniques
 We studied
 Transportation problem (Grand Prix)
 Assignment problem (Spring View)
 Logistics problem (RedBrand)
 Shortest path problem (Walk across states)
 Key takeaways
 Sparse representation
 Flow balancing equations
 Convert into network flow
Agenda

 Reminder: Course project


report/presentation due Dec 12
 Introducing integer programming
 Model setup is similar, but requires more
careful thoughts on logical constraints
 Examine business applications
 Capital budgeting
 Fixed cost manufacturing
 Set covering
 Facility location
 Production and transportation
Integer Programming

 When divisibility is not practical


 And integer value is a natural modeling
approach
 Example: funding either a project in full or not
funding the project
Capital Budgeting
Capital Budgeting – Tatham

 We would like to find an investment


portfolio that maximizes our total NPV.
 Investments do not allow partial participation
(i.e., all or nothing)

 Key issue: how to make sure the all or


nothing property of investment options
Model Setup

 Inputs:
 : index to represent different investment options

 : Cash requirement for investment , [e.g.,


 : NPV of investment , [e.g.,
 : Budget. [e.g.,
Model Setup (continued)

 Decisions:
 : Whether to invest in investment

 Objective:
 [maximize total NPV]

 Constraints:
 [total investment cannot exceed budget]
 [binary investment decision]
Implementation

Excel illustration.
Reflections on capital budgeting problem
 What if at most two projects can be
selected?

 What if investment 1 must be selected if 2


is selected?

 What if either investment 1 or 3 must be


selected?
Burrito Optimization Game - Gurobi

 https://www.gurobi.com/lp/academics
/burrito-optimization-game/
 Textile manufacturing process:
https://www.youtube.com/watch?v=YYWle
vX7Kw0
Textile Manufacturing at Great Threads

 Fixed cost in manufacturing


 Use Integer constraints

 Other than fixed cost, the problem is very


similar to a standard product mix problem
Model Setup

 Inputs:
 : index to represent different products

 : weekly rental cost to produce product type , [e.g.,

 : unit selling price of product type


 : unit variable cost to produce product type
 : total available labor hours. [e.g.,
 : total available sq. yd. of cloth. [e.g.,
Model Setup (continued)

 Decisions:
 : Whether to produce product
 : How many to produce product

 Objective:
 [maximize total profit]

 Constraints:
 [total labor hours cannot exceed available]

 [binary rental decision]


 [non-negative production]
Implementation

Excel illustration.
Reflections on fixed cost manufacturing
problem
 Key takeaway: incorporate all or nothing
fixed cost into the model through
constrained integer variables
The Set-Covering Problem
 Each member of a given set must be
covered by an acceptable member from
another set

 Airline hub and spoke system:


https://www.youtube.com/watch?v=mDkS
ehPF0M8
Hub Location at Western Airlines

 Determine the minimum number of hubs


to cover all cities

 A city is covered if it is within 1000 miles


of a hub
 So, whether a city is covered depends on
distance to hubs
 But we don’t know which cities are hubs
 Use integer variables to indicate whether a
city is a hub
Western Airlines Model Setup

 Inputs:
 : index to represent different cities

 Decisions:
 : Whether set city

 Objective:
 [minimize total number of hubs]

 Constraints:
 [each city must be covered]
Python illustration:
https://colab.research.google.com/drive/1UI7
261eRy_GCJqD-drWkrzH9vF2An_I5?usp=shar
ing
Reflections on set covering problem

 Is 1000 miles radius a good criteria?

 Key takeaway: use integer variable to


indicate whether a city is a hub, then
determine whether a city is covered
Locating and Assigning Service Centers at
United Copiers
Service Center Location at United Copiers

 Setting up three service centers while


minimizing total travel cost

 Different from Western Airline example,


here we know how many service centers
to set, but need to determine the best
location

 Use integer variables to indicate whether a


city should be a service center
United Copiers - Model Setup

 Inputs:
 : index of cities

 Decisions:
 : Whether to setup a service center at city
 : Whether to assign city to service center

 Objective:
 [minimize total trip distances]
Model Setup (continued)

 Constraints:
 [total number of service centers cannot exceed 3]
 [each city must be covered]

 [binary service center decision]


 [binary assignment decision]
Implementation

Excel illustration.
Reflections on facility location problem

 Why do we limit the number of service


centers to three?

 Key takeaway: use integer variable to


indicate whether a city has a service
center, then determine travel cost to
satisfy trip demand
Manufacturing and Distributing Fertilizer
Manufacturing and Distributing Fertilizer at
Green Grass

 Similar to United Copiers


 Financial metrics as opposed to travel
distances

 Use dual integer variables to determine


best production and shipment plan
Green Grass - Model Setup

 Inputs:
 : index of cities
Model Setup (continued)

 Decisions:
 : Whether to operate a plant in city
 : Whether to ship from plant to customer

 Objective: [maximize total profit]


 Constraints:
 [shipment cannot exceed plant capacity – if exist]

 [binary plant decision]


 [binary assignment decision]
Implementation

Excel illustration.
Reflections on production and transportation
problem
 Needs to determine both production and
transportation

 Key takeaway: use dual integer variables


to find the joint optimal solution
Summary

 Integer programming allows all or nothing


type of decisions
 Capital budgeting
 Fixed cost production
 Set covering
 Facility location
 Production and transportation
 Setup proper constraints to link integer
variables with non-integer variables to
enforce logical constraints

You might also like