CHAPTER 2

LITERATURE REVIEW

2.1 INTRODUCTION
Keeping in view the research need and in order to find the optimal solution for
a network of storm drainage system considering multiple objectives, literature review
was carried out, which is categorized into three subsections as rainfall-runoff
modeling, multiobjective, multicriteria decision making and its applications in water
sector, graph theory and graph metrics and its applications to network analysis. The
search was restricted to articles published in refereed journals related to the above
topics

2.2 RAINFALL-RUNOFF MODELLING

Literature review on this subtopic was further subdivided into an urban
drainage analysis and design, IDF curve generation, rainfall trends.

2.2.1 An Urban Drainage Analysis and Design

Conventional storm drainage design from hydraulic as well as hydrologic
analysis point of view is documented in detail by Gupta et al. (1983), Chow et al.
(1988) and Mays (2001).
The standard design constraints and methodology adopted for designing urban
storm drainage system according to the Indian standards can be referred in CPHEEO
manual on sewerage and sewage treatment documented by Ministry of Urban
Development, India (cpheeo.nic.in).
Use of legacy models, such as Storm Water Management Model (SWMM)
including the SWMM-RUNOFF and SWMM-EXTRAN blocks are engaged as the
rainfall-runoff and sewer models, respectively by number of researchers.
Mark et al. (2004) discussed the computer models of the drainage system such
as InfoWorks, MIKE URBAN, MOUSE, PCSWMM, and SWMM, which use St.
Venant’s one-dimensional equation for hydrodynamic modeling and generate
different scenarios, and understand the complex interaction between rainfall and

15
overflow. Further alternatives were evaluated and the optimal design solution was
found which could be implemented.
Maracineanu et al. (2009) used Hydraflow® tool extension for surface
drainage model design using three methods to define the design storm, Rational
Method, Modified Rational Method and Soil Conservancy System (SCS) Method as
well as three infiltration models: Horton Infiltration Equation, Green & Ampt
Algorithm, and SCS Infiltration Method. Using this software, surface water
management model was developed for catchment basins for drainage facilities and
storm sewer network design.
A real time storm sewer simulation system in combination with a SWMM
based on a pattern-oriented approach is demonstrated and studied by Lin et al. (2010).
Results helped to create a model for storm sewer hydrodynamics with hydraulic
facilities, thereby providing vital data such as stage and discharge for each junction
and pipe and aid in determining where the higher flooding regions are located based
on junction surcharge occurrence.
Details regarding the EPA SWMM which is a widely used program for
simulating urban runoff quantity and quality can be referred in user manual by
Rossman (2010). A user manual describes how to run the program. A reference
manual covers theory and algorithms, addresses common storm water management
issues, solution to design of problems encountered in practice, and provides additional
assistance in applying it to urban drainage design and analysis.
Similarly the use of Autodesk® SSA which is build upon SWMM platform in
storm water modeling is reported by Olson et al.(2010) elaborating how automation
innovations in Autodesk's newly released water resource engineering software aids
civil engineers to complete storm water analysis and design projects more efficiently
and economically.
Blumensaat et al.(2012) using modern Unmanned Air Vehicles (UAVs) used
high-resolution imperviousness maps for urban areas obtained as input in an urban
drainage modeling of a real life case study in the area of Lucerne, Switzerland.
Further, the influence of different types of images and their raster processing methods
on hydrological and hydraulic model performance was also evaluated and analyzed
for 307 individual sub-catchments for peak runoff and volume. The model predicted
channel flow was compared with flow measured at the catchment outlet.

16
 Integrated approach of runoff and sewer model using SWMM was studied by
Liao et al. (2012) which combined the conventional urban runoff and storm sewer
models with data transfer at runtime to simulate sewer flows with real time rainfalls.
A software system called IRES was developed which used the RUNOFF and the
EXTRAN block of SWMM as the legacy urban runoff and storm sewer models,
respectively. This study uses the Open Modeling Interface and Environment
(OpenMI) framework, which is the standard interface for model integration, as the
platform for dynamic data exchange in IRES.
Methodology used in Civil 3D® and its allied tools like Hydraflow® and
SSA®, an Autodesk® product, can be referred to in user manual of Autodesk®(2012)
for the detailed description.

2.2.2 IDF Curve Generation

Rainfall intensities of different frequencies and durations are the vital
parameters for the hydrologic design of storm water networks, culverts and many
other hydraulic structures. IDF relationships for various regions have been
constructed all across the world. Also, the geographical distribution of IDF
relationships has been studied in several developed countries and maps have been
developed to provide the rainfall intensities or depths for various return periods and
durations which can be referenced in standard references like Subramanya (1984),
Chow et al. (1988) and Viessman et al. (1989).
Rambabu et al. (1979) developed an equation for IDF relationship for Indian
region analyzing rainfall characteristics for 42 stations.
Kothyari and Garde (1992) developed a general relationship for IDF by
analyzing data of 80 raingauge stations in India. The correctness of form of the
developed relationship was confirmed by its comparison with existing equations and
use of data from different hydrological regions. The applicability of this relationship
to the data from other countries was verified with the limited available data set.
Koutsoyiannis et al. (1998) proposed a new comprehensive approach to the
formulation of IDF curves using efficient parameterization. Several appropriate
distribution functions were studied for that purpose and a simple analytical
approximation of the most common distribution functions were documented which
allow expressing IDF relationships in a mathematical form. A major part of Greece

17
was considered for the analysis to document the methodology for the construction and
regionalization of the IDF relationship.
Gerold and Watkins (2005) worked out IDF estimates as important statistical
precipitation records used for hydrologic engineering design and updated it for the
state of Michigan USA. They applied a regional frequency analysis procedure and the
simple scaling model of short duration data of less than one hour available in the
Detroit metropolitan area.
Nhat et al. (2006) established IDF curves and derived generalized IDF
equation for the seven stations in Vietnam. The parameters of rainfall IDF equations
were developed for ungauged areas to estimate rainfall intensity for various return
period and rainfall duration. The parameter contour maps were prepared to estimate
ungauged rainfall with different return periods.
Guo (2006) elucidated the importance of modifying IDF curves to design
storm drainage system. In this study, the rainfall IDF relationships were determined;
from previous analysis results for Chicago city, to depict the climate conditions of the
first and second halves of the last century. Using these IDF relationships, the impact
of the observed increment in heavy rainfall events on the design and performance of
urban drainage systems was estimated. Results of the estimate emphasize the need for
updating rainfall IDF relationships to reflect altering climate conditions and its effect
on the design of new as well as replacement of an old urban drainage system.
Raiford et al. (2007) updated the existing IDF curves in the region of United
States, obtained these curves at ungauged sites throughout the region using the newly
developed rainfall frequency analysis technique and developed the depth-duration-
frequency curves as well as isopluvial maps.
Chowdhury et al. (2007) studied the maximum rainfall area in Bangladesh,
named Sylhet. Indian Meteorological Department (IMD) empirical reduction formula
was applied to estimate the short duration rainfall intensity from daily rainfall data.
Analysis carried out using various distribution functions and Chi-square goodness of
fit test applied to identify the best statistical distribution among them.
El-Sayed (2011) used rainfall frequency analysis technique to ungauged sites
in Egypt with six different durations varying from 5 minutes to 24 hours for various
return periods and the IDF curves and isopluvial maps for the region were developed
using the available rainfall data.

18
2.2.3 Rainfall Trends
Exhaustive literature related to the trend detection techniques is available in
hydrological and environmental field.
Work concerning the Spearman rank correlation test carried out by
Lettenmaier (1976) which considers a method developed for summarizing the power
of the parametric tests, the non-parametric Spearman's Rho test and Mann-Whitney's
test against step and linear trends in a dimensionless trend number.
Some of those studies include least squares linear regression for the detection
of trends in time series of hydrological variables by Haan (1977).
Seasonal Mann–Kendall test was applied for detection of the trend which can
be referred to Helsel and Hirsch (1992).
Trend analysis of rainfall and temperature data for India was studied by Jain
and Kumar (2012).
Sonali and Kumar (2013) describe various tests for detecting trends like Sen’s
slope and other innovative trend analysis techniques. The study performed the spatial
and temporal trend analysis of annual, monthly, seasonal maximum and minimum
temperatures in India, using variety of non-parametric methods which considered the
effect of serial correlation during analysis.

2.3 MULTIOBJECTIVE, MULTICRITERIA DECISION MAKING, AND

ITS APPLICATIONS IN WATER SECTOR.
Optimal design of storm water network is studied by various researchers either
focusing on explicit design of horizontal layout and vertical layout fulfilling multiple
objectives separately or by implicit simultaneous design.
Mays and Wenzel (1976) studied optimization of vertical layouts, on branched
sewer system using dynamic programming.
Afshar (2010) described the application of the continuous Ant Colony
Optimization Algorithm (ACO) to optimal design of sewer networks. Two alternative
approaches to implement the algorithm were presented and applied to a storm sewer
network. The decision variables considered were the node elevations of the network
of the optimization problem. Two alternative formulations of the constrained
algorithm were used to solve a test example and the results were compared with those
of unconstrained approach. The methods were reported as more effective in locating

19
the optimal solution and efficient in terms of the convergence characteristics of the
resulting algorithms.
General framework for risk based optimal storm drainage design can be
referred to Sun et al. (2011). The decision-making process for storm sewer network
design considering flood risk was explored in this study. A flood risk based storm
sewer network design method incorporating a multiobjective optimization and a
choice process is developed with different design criteria. The storm sewer network
design considering flood risk was formed as a single objective optimization problem
with condition that the decision criterion was given a priori. A framework for this
approach with a single objective optimization was developed with GA as the
optimizer.
Haghighi and Bakhshipour (2012) focused on use of the nonlinear and discrete
constraints of an optimization model to design sewer networks. They have developed
an adaptive Genetic Algorithm (GA) wherein each chromosome, consisting of sewer
diameters and slopes and pump indicators, is a feasible design. Generation of the
binary chromosomes and then decoding it to feasible design alternatives followed by a
sequential design analysis and algorithm implementation was carried out. The
adaptive decoding strategy was set up based on the open channel hydraulics and
sewer design criteria satisfying all the sewer system constraints. Proposed approach
was then applied to a benchmark sewer network from the literature
Palumbo et al. (2014) proposed a general procedure for the optimal design of
urban drainage networks taking into account practical technical constraints and the
management cost of the automatic flushing devices. The procedure utilized a standard
genetic algorithm, tested on two case studies with reference to an urban drainage
network taken from the literature to evaluate the impact of different input parameters
and constraints on the optimization problem. Effect of crown elevation of the network
as a variable in a given range, the effect of constraint of size of downstream pipe, and
the effect of modifying the peak wastewater discharge or the management unit costs
were considered.
Simultaneous optimization of horizontal as well as vertical layouts by
developing an iterative model has been studied by number of researchers. This
includes the design of urban drainage systems that consists of solving two sub
problems: the first, generating the layout and second sizing the sewers, which includes
pipe diameters, installation depths, and pumping station.

20
 A new approach for the optimization of urban drainage systems including the
optimal selection of layout using the searching direction method and the optimal
design of a given layout by discrete differential dynamic programming was developed
and applied in a case study by Li and Matthew (1990). It was shown that,
consideration of hydraulic factors such as flow rate, sizes, gradients of pipes, and the
effect of online pumping stations is very important in urban drainage systems along
with optimal layout selection, and can achieve much greater construction cost savings
than that without considering layout optimization. The study highlighted that the
minimum buried depth design is not the global optimum design but the global
optimum design is the alternative of the optimal balance amongst the buried depth and
the number and locations of online pumping stations.
Diogo et al. (2000) developed a global optimization model for simultaneously
obtaining the optimal layout and design of urban drainage systems for wastewater and
stormwater. The model included the optimal location for pumping systems and
outfalls or wastewater treatment plants and allowed the presence of special structures
and existing subsystems for optimal expansion. Two basic optimization levels were
considered; first the generation and transformation of general layouts of forests of
trees until a convergence criterion is reached, and in the second, the design and
evaluation of each forest. The methodology combined and developed a sequence of
optimal design and plan layout sub-problems. Robust Dynamic programming was
used along with Simulated Annealing (SA) and GA, in this discrete optimization
problem of higher dimension.
Haghighi (2013) described a method to solve the layout sub problem based on
the graph theory. All possible sewer lines and manholes were considered for the sewer
network so that an undirected looped graph was formed at the start for a given city.
This was followed by an adaptive procedure called loop-by-loop cutting algorithm
developed to create feasible sewer layouts from the base graph. Using this algorithm,
all constraints of the sewer layout sub problem were systematically handled. By
defining an objective function, the optimum layout was achieved by employing a
simple genetic algorithm. After the sewer layout was determined, pipes and pumps
were designed by using a discrete differential dynamic programming model. A
benchmark sewer system was also considered for optimization from the literature.
There are different methods for multicriteria decision, which could be
categorized as distance based, outranking based, utility based and mixed category

21
based methods. Some of these methods are Composite Programming, Compromise
Programming (CP), Technique for Order Preference by Similarity to Ideal Solution
(TOPSIS), Analytic Hierarchy Process (AHP), ELimination and Choice Expressing
REality (ELECTRE) I-III, Preference Ranking Organization METHod for Enrichment
of Evaluations (PROMETHEE) methods, details of which, can be referred to Zeleny
(1982) and guidelines for selection of appropriate method can be studied from
Guitouni and Martel (1998).
For peak discharge control, though planning and design procedures are
available, still there has been a need for similar methods for the control of channel
erosion and water quality. Thus, a procedure that could be used for design was
provided along with a relationship for making planning estimates of the volume of
detention storage required to control channel erosion with reference to study area by
McCuen and Moglen (1988). They also highlighted the importance of adopting
multicriteria storm-water management policies, with the policy recommending control
of flooding, channel erosion, and detention time for water quality enhancement.
Zadeh (1988) described some basic ideas underlying fuzzy logic and some
representative applications. The author covered basic principles; meaning
representation and basic rules of inference; as well as the linguistic variable and its
application to fuzzy control.
Tools developed for the design of urban storm drainage networks allow
multiple solutions to be proposed for the same problem and then the decision maker
has to choose which solution to adopt. At the same time, multiobjective decision-
making is now directed from optimization methods to more interactive decision
support tools. To arrive at feasible choices of alternative techniques like infiltration
and detention systems in urban storm-water drainage project, Azzout et al. (1995)
formalized the decision-making process, taking account of the multiple criteria and
opinion of the large number of partners involved.
A fuzzy compromise approach to decision analysis related to water resource
systems planning under uncertainty was described by Bender and Simonovic (2000).
The study compares the ELECTRE method with the fuzzy compromise approach. The
comparison highlights the benefits of adopting a multicriteria decision analysis
technique, which presents subjectivity within its proper context at the same time
maintaining an intuitive and transparent technique for ranking of the alternatives. The
fuzzy compromise approach allowed a family of possible conditions to be reviewed,

22
and support group decisions through fuzzy sets designed to reflect collective opinions,
conflicting judgments, and the effect of risk tolerance differences among decision
makers.
Kolsky and Butler (2002) described conceptual and practical aspects of urban
storm drainage performance indicators, particularly for India. They have elaborated a
general framework of objectives and performance indicators as logical intermediate
steps between values and the decisions taken to reflect them, followed by selecting
practical approaches to performance and indicator measurement, built upon field
experience in India.
Assessing long-term, cost-effective drainage options, numbers of factors are to
be considered including technical, environmental, social, and economic cost. This was
taken into account by Ellis et al. (2004) who had developed a multicriteria analysis
methodology for the evaluation and accreditation of the selection of sustainable
drainage structures within the limits of an overall decision-support framework
supporting benchmark standards.
Blanpain et al. (2004) presented a tool, which could be applied to design of an
urban storm drainage networks. This allowed several solutions to be proposed for the
same problem to fulfill criteria like economic, ecological, functionality and
maintenance. This tool integrated multicriteria, fuzzy sets and the base of rules
concepts. A method, which helped to evaluate one of the criteria related to the global
functioning of a storm drainage network, was defined, tested, validated on imaginary
cases, and consequently applied to an actual case of re-organization of an urban storm
drainage network of the city of Annequin, France.
Abrishamchi et al. (2005) successfully implemented MCDA methods in water
resource planning and management. CP method was applied to aid decision makers in
selecting the most preferred alternative from water distribution network solutions.
This network was used for distribution of both available and the transmitted water in
the Zahidan city in Iran to satisfy future water demands by a long-distance water
transmission project. The results indicated that decision-makers for comprehensive
urban water management studies could employ the method.
Hajkowicz and Collins (2007) reviewed 113 published water management
multicriteria analysis studies from 34 countries. They found that multicriteria analysis
is being used for evaluation of water policy, to plan water management projects
strategically, and selecting the best infrastructure solution. A wide range of

23
multicriteria analysis methods are being used with the fuzzy set analysis, paired
comparison and outranking methods being most common. The paper also examined
the motivations for adopting multiple criteria analysis in water management problems
including future research directions.
Researchers like Marinoni et al. (2009) describe a decision support software
system referred to as the Multiple Criteria Analysis Tool (MCAT) and also utilized
CP method. MCAT identifies a portfolio of decision options that return a maximum
aggregated benefit under a constrained budget. Primarily, the tool was developed to
be used in natural resource management contexts illustrated through three natural
resource management case studies in Australia.
Khatri et al. (2011) also applied fuzzy set theory to urban infrastructure
system. This paper documents a framework for characterizing, analyzing, and
computing the performance of urban infrastructure systems based on a systems
approach. A fuzzy synthetic evaluation technique was developed to synthesize
performance indicators into an index, which was generated by adding the performance
indicators in the first three hierarchical levels. The overall performance of the urban
infrastructure systems then was documented by merging the indexes of individual
infrastructure systems. The relative importance of the performance indicators during
the synthesis process is determined using the AHP method. The framework depicted
the performance of the civil infrastructure systems in Kathmandu, Nepal to help
decision makers in identifying the weak and strong components of urban
infrastructure systems and to formulate apt strategies for infrastructure planning based
on performance and investments.
AHP as described by Saaty (1990) has been applied by researchers in various
application areas. The paper describes the general decision operation by decomposing
a complex problem into a multi‐level hierarchical structure of objectives, criteria,
subcriteria, and alternatives. The AHP provides a fundamental scale of absolute
magnitudes to represent judgments in the form of paired comparisons. A ratio scale of
relative magnitudes expressed in priority units is then derived from each set of
comparisons. An overall ratio scale of priorities is synthesized to obtain ranking of the
alternatives. An application of the AHP to a retributive ongoing conflict in which the
parties maximize both their benefits and costs to the opponent was presented in this
paper.

24
 Lai et al. (2002) applied AHP related to multimedia authorization system
selection in a group decision environment.
Ackerman and Stein (2008) applied five multicriteria decision making
methods and finalized alternatives using AHP to the criteria such as volume and the
peak flow rate reductions, pollutant loads and best management practices.
AHP was applied to water resources planning by Cabrera et al. (2011)
highlighting how to make use of the AHP technique to improve the decision-making
process. This helped to satisfy new water demands in a local context considering new
technologies to be provided in the form of sustainable solutions meeting future water
demands.
Siraj et al. (2015) developed an interactive decision support tool ‘PriEsT’ to
estimate priorities from pairwise comparison judgments, based on AHP. The tool can
aid in validating results of AHP.
Cai et al. (2004) successfully implemented application of group decision
theory. They proposed a method based on compound models for regional water
resources planning. This method combined modeling techniques such as
multiobjective analysis, multicriteria analysis involving multiple decision makers, and
supported plan generation and evaluation, individual and group preference elicitation,
negotiation to arrive at a consensus plan.
TOPSIS is a distance based MCDA technique that was first proposed by
Hwang et al. (1993). They proposed TOPSIS for multiobjective decision-making
algorithm for solving multiple objective decision-making problems by considering
two reference points of the positive ideal solution and the negative ideal solution
simultaneously. Thus, an incommensurable and conflicting objective function was
reduced into auxiliary bi-objective commensurable but conflicting objective function.
This was followed by use of the max-min operator, obtaining a compromise solution
to the problem.
TOPSIS has been used in design, engineering, environmental and water
resources management because of its simplicity. Afshar et al.(2011) applied fuzzy
TOPSIS model to provide a framework to help water managers identify critical issues,
attach relative priorities to those issues, select best compromise alternatives from an
abundance of project purposes and resource values, subjected to conflicting policy,
planning, and management decisions related to a real water resource management
problem in Iran. Results in the form of ranks of the alternatives were presented using

25
deterministic and fuzzy based model, which highlighted that the model may be used
in a large-scale multi-level assessment process.
State of the art literature review to classify and interpret the ongoing and
emerging issues that apply the TOPSIS methodology has been carried out by
Behzadian et al. (2012). The paper includes a detailed literature survey to taxonomize
the research on TOPSIS applications and methodologies. The researchers have
reviewed 266 scholarly papers from 103 journals since the year 2000, separating them
into different application areas like health, safety, environment management, energy
management and water resources management along with recommendations for future
research in TOPSIS decision-making.

2.4 GRAPH THEORY AND GRAPH METRICS RELATED TO

NETWORK ANALYSIS.
Tekeli and Belkaya (1986) focused on computerized horizontal layouts of
network generation for sanitary sewers. A layout generation algorithm was developed
for generation of sanitary sewer layouts, using a common shortest path algorithm. For
selection of the shortest path from each manhole to the final outlet, three sewer cost
attributes were formulated considering surface elevations at the manholes and sewer
lengths. From the minimum cover depth and sewer slope volume of excavation was
calculated. This led to the minimal excavations and invert depths when the generated
layouts were hydraulically analyzed and designed. The results were satisfactory for
the networks with up to 70 manholes. In addition, the guidelines were developed for
sub-zoning as well as general implementation by testing of the algorithm with various
sized networks.
Walters and Smith (1995) applied evolutionary algorithms to optimize tree
type networks, which occur in sewerage, irrigation, water, gas supply and distribution,
and in many other engineering applications. The model was based upon evolutionary
design and GA principles, which include the use of an efficient tree-growing
algorithm and the incorporation of redundant ‘genetic’ information within the
‘reproduction’ phase. The model was tested on two test cases.
Dunn and Wilkinson (2012) applied graph theory metrics to network flow
models, to identify critical areas within infrastructure systems. A reduced complexity
flow model that could be used to simulate flows in infrastructure networks was
developed to simplify the physical problem and therefore produce the most applicable

26
range of infrastructure networks. An initial estimation of the applicability of graph
theory metrics to infrastructure networks was made by comparing the distribution of
flows, calculated using this model, to the shortest average path length in three of the
most recognized classes of network: scale-free networks, small-world networks, and
random graph models. Further, they have demonstrated that for all three classes of
network there is a strong correlation. In addition, the efficacy of graph theory metrics
to predict and identify susceptible areas in a specific type of infrastructure system was
documented.
Lhomme et al. (2013) applied network topology theory to analyze, design, and
manage infrastructure systems. The methodology designed focused on a system
approach emphasizing technical network’s importance concerning urban resilience
and how urban networks are able to face natural hazards. A Web-geographical
information system was developed for analyzing resistance capacity, absorption
capacity and recovery capacity of different technical networks and the developed
interface was applied on a dense road network in France.
Infrastructure systems including water networks have resilience as a common
objective. In the study of complex systems, the structure and function of networks can
contribute directly to system resilience. Network theory, describes how connectivity
between individual components can explain some system-wide properties of expanse
and reliability. Water resource system analysis was carried out by applying techniques
from network theory and complexity science to assess adaptability and resilience by
Porse and Lund (2015). In this paper, an analysis of connectivity in a network model
of California's water infrastructure system using several network science techniques
indicated that nodes in California's water system are clustered but without scale-free
properties. It is also illustrated that how adaptability, not just complexity, influences
resilience in planning of the civil infrastructure networks.
Bastian et al. (2009).developed Gephi®, free open source codes and tools in
order to calculate various network related parameters like Characteristic Path Length
(CPL), diameter, stress etc.
CentiScaPe® plug-in developed by Scardoni et al. (2009) was used to generate
various network centrality measures. The tool also allowed the user to analyze
existing relationships between experimental data provided by the users and node
centrality values computed by the plug-in. CentiScaPe® identifies network nodes that
are important from both experimental and topological viewpoints. CentiScaPe® also

27
provides a Boolean logic-based tool that allows easy characterization of nodes whose
topological relevance depends on more than one centrality.
In addition, SANET tool developed by Okabe et al. (2006) and
COMPLEXNETGISV02 developed by Caschili (2010) that are available on
ARCGIS® platform, help compute various network metrics.
Generating a spanning tree is a well-known problem in the area of network
optimization. A spanning tree is a connected acyclic sub-graph that spans all nodes.
There are various applications of finding degree constrained spanning tree, namely in
the design of road systems, where a limit is imposed on the number of roads that are
allowed to meet at any crossing. This generalized problem has been tackled by
Savelsbergh and Volgenant (1985), describe a branch and bound algorithm to solve
this problem and show the results of calculation for problems up to 200 vertices.
In the design of computer communication networks, as well as in designing
VLSI layouts, electrical circuits, the overall design problem was formulated as a
nonlinear combinatorial optimization problem of topological design of computer
networks by Gavish (1982). Problem involved decisions on where to place Network
Control Processors (NCPs), selecting the set of backbone links to connect NCPs,
linking end user nodes to the NCPs, and deciding on the set of routes which support
communications between communicating end user node pairs. A Lagrangean
relaxation of the problem was presented and effective solution procedures of the
Lagrangean problem was developed which provides lower bounds on the optimal
solutions to the complete problem. The Lagrangean-based solutions were further
improved using subgradient optimization procedures and three heuristics were
developed for generating feasible solutions with restriction of 200 end user nodes and
maximum 30 NCP locations.
Minieka (1978) documented Dijkstra’s method of graph theory to generate the
shortest spanning tree of a graph.
The details regarding network measures can be referred to Kansky and
Danscoine (1989) as well as the details regarding network flows and related theory
can be referred to Ahuja et al. (1993).

2.5 OBSERVATIONS BASED ON LITERATURE REVIEW

Based on literature reviewed in three sections, following critical observations
are deduced.

28
2.5.1 Observations Based on Literature Review on Rainfall-Runoff Modeling
In case of analysis of rainfall, data extensive work has been done pertaining to
disaggregation, fitting data in to probability distribution and generating IDF curves or
update IDF curves. However, very few contributions are available on study area i.e.
Pune city area and even though IDF curves are generated for the study area, these are
not updated and fitted using various probability distributions. In fact, results of IDF
curve using step method proposed by CPHEEO and only Gumbel’s Extreme value
distribution were compared to arrive at IDF curves. Also as the city is developing and
spreading on a vast area, there is a need for assessing variation in rainfall pattern and
associated variation in IDF curves was felt.
After reviewing papers on hydraulic and hydrological analysis of storm
drainage system, it was found that number of researchers have incorporated use of
available commercial tools like Hydraflow, Storm CAD®, XPSWMM®, PCSWMM®,
and INFOSWMM® for which integrate Geographical Information System (GIS) data
with analysis and built-in SWWM tool. With Autodesk® Civil 3D® software which
integrates Computer Aided Drafting (CAD), GIS and recently included analysis tools
like Hydraflow® and SSA®, makes it one of the most effective tool. The tool is being
used by most professionals in design engineering and educational license is made
available to researchers. Thus to avail the capabilities of the tool, a methodology was
required to be documented to generate near optimal solutions, which would further
lead to selecting the best solution.

2.5.2 Observations Based on Literature Review on Multiobjective,

Multicriteria Decision Making and its Applications
Exhaustive literature review on multiobjective and multicriteria optimization
techniques reveals that number of heuristic techniques like GA, SA, ACO have been
applied to water or wastewater network in search of global optimal solution obtained
from a set of continuous set of Pareto optimal solutions. The methods applied are
expensive in terms of time and do not assumptions during modeling related to
explicit or implicit constraints. In addition, implementation of these methods on large
networks can be very time consuming. On the other hand, MCDA methods are
applied on set of discrete solutions and are less time consuming and pose more
effective and simplistic methodology, if applied on the set of near optimal solutions or

29
alternatives obtained using system approach in urban storm drainage design. MCDA
has been implemented in urban storm drainage design with more focus on
sustainability and water quality worldwide. From Indian perspective, although these
issues are very vital, less attention is given to these by the engineers as well as
decision makers (CPHEEO manual). Reasons may be attributed to increase in cost of
network, improper planning of land development leading to irregularly planned road
and utility networks, difficulty in maintenance of such structures, encroachments on
natural channels and waterways hampering the water quality. Also very few
researchers have considered criteria related design and execution, network topology
and performance of the network using MCDA to arrive at the most preferred solution.

2.5.3 Observations Based on Literature Review on Graph Theory and Graph

Metrics Related to Network Analysis
Concepts from graph theory and graph metrics have been utilized in most of
the networks whether it is road network, electrical lines, pipe networks, complex
biological networks, or social network. The network optimization techniques have
been successfully implemented on pipe networks, more specifically water distribution
networks where in, number of layouts are generated and analyzed structurally as well
as functionally. A very few research contributions are made on directed pipe networks
like gravity based storm water pipe networks with varying network layouts either tree
or closed and effect of layout on hydraulic performance as well as costs.
Thus, need was felt to develop a methodology which will generate a set of
near optimal solution of storm drainage network, and then by applying MCDA
techniques to arrive at the best solution from the given set. Similarly, an algorithm
and tool could be developed which will help storm drainage designers to generate
different layouts considering constraints of roads, terrain slopes, and discharge point
location.

2.6 CLOSING REMARKS

From the reviewed literature, it is revealed that exhaustive work has been
carried out on specific regions of countries worldwide, related to precipitation
analysis. MCDA methods have been applied to sustainable storm drainage system and
optimization of wastewater networks considering number of heuristic techniques. As
the region changes, objectives as well as their relative importance or weights differ

30
and also the level of acceptance of the designed solutions. Thus different methods are
required to be applied and tested in different regions in order to obtain the most
preferred and acceptable solution specifically in urban infrastructure development
from Indian context.
Considering the key areas where research input is required, relevant data in
either primary or secondary form was required to be procured and processed so that it
can be further analyzed and utilized in design to arrive at an optimal solution for an
urban storm water drainage system.
The next chapter deals with the details of the study area selected and spatial
and non spatial data collected and preprocessed in order to analyze and design an
urban storm drainage system.

31