This document contains abstracts for 7 doctoral dissertations submitted for the 2000 Transportation Science Section Dissertation Prize Competition. Two dissertations were awarded First Prize. Hillel Bar-Gera of the University of Illinois at Chicago developed origin-based algorithms for transportation network modeling that provide accurate solutions using less memory than traditional approaches. Diego Klabjan of the Georgia Institute of Technology addressed topics in airline crew scheduling and developed parallel computing techniques to solve large optimization problems in airline operations. The remaining 5 dissertations made valuable contributions to transportation science but were not given honorable mentions.

1. This article was downloaded by: [54.162.69.248] On: 20 May 2020, At: 10:48
Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
INFORMS is located in Maryland, USA
Transportation Science
Publication details, including instructions for authors and subscription information:
http:/ / pubsonline.informs.org
Abstracts for the 2000 Transportation Science Section
Dissertation Prize Competition
Paolo Toth, Hillel Bar-Gera, Diego Klabj an, Roberto Baldacci, Cenk Caliskan, Christina
Diakaki, Lasse Fridström, Vij ay S. Nori,
To cite this article:
Paolo Toth, Hillel Bar-Gera, Diego Klabj an, Roberto Baldacci, Cenk Caliskan, Christina Diakaki, Lasse Fridström, Vij ay S.
Nori, (2001) Abstracts for the 2000 Transportation Science Section Dissertation Prize Competition. Transportation Science
35(1):99-103. https:/ / doi.org/ 10.1287/ trsc.35.1.99.10139
Full terms and conditions of use: https://pubsonline.informs.org/Publications/Librarians-Portal/PubsOnLine-Terms-and-
Conditions
This article may be used only for the purposes of research, teaching, and/ or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact permissions@informs.org.
The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.
© 2001 INFORMS
Please scroll down for article—
it is on subsequent pages
With 12,500 members from nearly 90 countries, INFORMS is the largest international association of operations research (O.R.)
and analytics professionals and students. INFORMS provides unique networking and learning opportunities for individual
professionals, and organizations of all types and sizes, to better understand and use O.R. and analytics tools and methods to
transform strategic visions and achieve better outcomes.
For more information on INFORMS, its publications, membership, or meetings visit http:/ / www.informs.org

2. Abstracts for the 2000 Transportation
Science Section Dissertation
Prize Competition
Paolo Toth
DEIS, University of Bologna,
Chair, 2000 TSS Dissertation Prize Competition
The TSS Dissertation Prize Competition is the old-
est and most prestigious award a young scholar in the
Transportation Science area can receive. On behalf of
the TSS Dissertation Prize Committee, I am pleased
to provide the report on the 2000 competition.
Eligible doctoral dissertations were those com-
pleted and submitted between June 1, 1999, and May
31, 2000, in the general area of Transportation Sci-
ence. To be considered, a dissertation had to be nomi-
nated by the thesis supervisor. The competition draws
entries from the top transportation programs around
the world. The committee received seven submis-
sions, each of which was carefully reviewed by the
prize committee.
The members of the committee were:
• Professor Ismail Chabini, Massachusetts Institute
of Technology
• Professor Oli Madsen, Technical University of
Denmark
• Professor Pitu Mirchandani, University of Ari-
zona
• Professor Jean-Yves Potvin, University of Mon-
treal
I sincerely thank all the members of the committee
for the time and effort they devoted to the difficult
task of judging the theses. The committee provided
thoughtful and timely reviews and was a very colle-
gial group. It was really a pleasure to work with them.
The quality of all the theses submitted was excel-
lent, and the choices were very difficult. This is a clear
indication of the extraordinary level and productiv-
ity of the field of Transportation Science. The com-
mittee unanimously felt that two theses were really
outstanding and deserving the First Prize. As for the
remaining five entries, all deserved honorable men-
tions, but it was impossible for us to choose only some
of them, so we decided to give no honorable mention
this year. The First Prizes were awarded to Dr. Hillel
Bar-Gera of the University of Illinois at Chicago,
and to Dr. Diego Klabjan of the Georgia Institute of
Technology. The supervisors were Prof. David Boyce
and Prof. George L. Nemhauser, respectively. The
awards were announced at the INFORMS meeting in
San Antonio in November 2000.
Abstracts of the seven submissions are provided
below, with the two winners first and the remain-
ing submissions in alphabetical order. Each makes a
unique and valuable contribution to transportation
science, and all entrants and their supervisors deserve
warmest congratulations.
FIRST PRIZE
Origin-based Algorithms for Transportation
Network Modeling
Hillel Bar-Gera
The transportation planning process relies on travel
forecasts that result from various transportation mod-
els. Some of the well-known models are formulated
as nonlinear convex optimization problems. Solv-
ing these problems is quite challenging due to their
nonlinear nature and their combinatorial structure.
Large-scale networks of practical interest increase the
need for computationally efficient algorithms.
At the heart of most transportation models stands
the traffic assignment problem, which is to predict the
0041-1655/01/3501/0099$05.00
1526–5447 electronic ISSN
Transportation Science © 2001 INFORMS
Vol. 35, No. 1, February 2001 pp. 99–103

3. DISSERTATION ABSTRACTS
route choice of travelers given the origin and destina-
tion of each traveler, under the assumption that each
traveler seeks to minimize the time/cost associated
with his chosen route. Almost all known algorithms
for the traffic assignment problem are either link-
based or route-based. Link-based algorithms fail to
achieve highly accurate solutions, while route-based
algorithms are regarded as impractical for large-scale
networks due to their large memory requirements.
This research presents a different approach in
which the solution is represented by origin-based link
flows aggregated over all destinations. This approach
provides highly accurate detailed solutions; its mem-
ory requirements are reasonable even for large-scale
networks, while computation times are substantially
lower than for link-based algorithms.
Formal mathematical definition of the proposed
algorithm for the traffic assignment problem is pre-
sented, followed by a proof of its convergence. Com-
putational results on several networks, including
a large-scale one, demonstrate the superior perfor-
mance of the proposed algorithm.
A direct procedure to extract route flows from
origin-based solutions is presented, based on a
behavioral assumption named bypass proportional-
ity. The close relationship between this assumption
and entropy maximization is examined. The ability
to disaggregate origin-based link flows to route flows
demonstrates that the detail provided by origin-based
solutions is practically equivalent to the detail pro-
vided by route-based solutions, while the memory
requirements of origin-based solutions are substan-
tially lower.
This thesis was submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in Civil
Engineering at the University of Illinois at Chicago. The
thesis supervisor was David Boyce.
FIRST PRIZE
Topics in Airline Crew Scheduling and
Large Optimization
Diego Klabjan
Airline operations research is a well-studied area
due to its large impact on revenue. The problems
of assigning equipment types, aircraft and crew
itineraries to flights are crucial, since a slight improve-
ment in a solution value can yield millions of dollars
in savings. The focus of this dissertation is to improve
on existing solution methodologies in planning of an
airline and to partially integrate the assignment prob-
lems cited above.
The main topic of the dissertation is airline crew
scheduling, the problem of assigning crew itineraries
to flights. We develop a new methodology for solv-
ing crew scheduling problems. The approach uses
parallel computing techniques and algorithms and
hence is implemented in parallel and run on a clus-
ter of personal computers. We present a parallel ran-
dom itinerary generator and a parallel solver for large
scale linear programs. The results obtained are very
encouraging.
We develop a parallel primal-dual simplex algo-
rithm that is capable of solving linear programs with
at least an order of magnitude more columns than
the previous work. The algorithm repeatedly solves
several linear programs in parallel and combines the
dual solutions to obtain a new dual feasible solution.
The primal part of the algorithm involves a new ran-
domized pricing strategy. We tested the algorithm on
instances with thousands of rows and tens of millions
of columns.
Part of the dissertation deals with partially integrat-
ing schedule planning, aircraft assignment, and crew
scheduling. We solve a crew scheduling problem with
additional constraints that preserve the plane count,
and we assume we are allowed to change the depar-
ture times slightly. We improve on the existing crew
scheduling solutions, on average, by a factor of two.
The last part of the dissertation focuses on a new
methodology for solving the airline crew scheduling
problem over the weekly horizon. We introduce a sec-
ond objective, called regularity. Regularity measures
the repetition rate of crew itineraries. New models are
developed that address the cost of a crew schedule
and the regularity. The models produce huge integer
programs, and we present a methodology for solving
them. Solutions are reported with low cost and high
regularity.
This thesis was submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy
in Algorithms, Combinatorics and Optimization at the
Georgia Institute of Technology. The thesis supervisor was
George L. Nemhauser.
100 Transportation Science/Vol. 35, No. 1, February 2001

4. DISSERTATION ABSTRACTS
Algorithms for Location and Routing Problems in
Distribution Systems
Roberto Baldacci
This thesis is concerned with the study of exact and
heuristic algorithms for location and routing prob-
lems in distribution systems. We have considered
the Capacitated p-Median Problem (CPMP), the Vehi-
cle Routing Problem (VRP), the VRP with backhauls
(VRPB), the Traveling Salesman Problem (TSP) with
mixed deliveries and collections, and the TSP with
backhauls.
We have chosen the Set Partitioning (SP) approach
to model both CPMP and VRPB, since it can be
extended easily to deal with a wide range of con-
straints without changing the structure of the model.
We propose an additive method for computing a valid
lower bound that finds a feasible solution of the dual
of the LP-relaxation of SP without generating the
entire SP matrix. The dual solution obtained is used
to reduce the SP variables, and the resulting problem
is then solved by a branch and bound algorithm. The
computational results show that the new exact algo-
rithms for the CPMP and the VRPB are able to solve
instances of size up to 100 customers.
As part of this research, we extended the two
commodity network flow formulation of the TSP
proposed by Finke et al. (1984) to derive new inte-
ger formulations for routing problems like the TSP
with mixed deliveries and collections, the TSP with
Backhauls and the VRP. These formulations are used
to derive new lower bounds based on linear relax-
ation strengthened by new valid inequalities. The
computational results show that the branch and cut
algorithms for the two variants of the TSP are able
to solve to optimality problems involving up to 150
customers, while the exact algorithm for the VRP can
solve problems with up to 100 customers.
With reference to heuristic methods, we designed
a Bionomic Algorithm for the CPMP with additional
constraints. The computational results show the effec-
tiveness of the proposed approach, when compared
to the best performing heuristics presented in the
literature.
Finally, a real-life problem of resource planning in
the utility sector is approached as a Multi-Depot-
Period-VRP, and a heuristic algorithm has been devel-
oped for it. The computational implementation of the
model is described, and results are obtained for a spe-
cific case study.
This thesis was submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in
Operational Research at the University of London. The the-
sis supervisor was Eleni Hadjiconstantinou.
Empty Equipment Allocation and Vehicle
Scheduling for Long-Haul Networks
Under Cyclic Constraints
Cenk Caliskan
The main objective of this research is to develop
realistic yet accurate and efficient models for the
empty equipment reallocation and crew scheduling
problem encountered in the long-haul trucking indus-
try. The problem can be defined as follows: given a
set of origin-destination demands for movements of
trailer-loads between consolidation centers, optimize
the movements of trailers, tractors and drivers while
ensuring that drivers return home within a predeter-
mined period of time. We study the problem within a
less-than-truckload (LTL) environment, but the results
are applicable to all long-haul networks.
In the first part of this research, we cast the prob-
lem in a tactical framework and develop a static linear
programming model that optimizes the average flows
of tractors and trailers given: demand constraints
between pairs of consolidation centers, conservation
of flow constraints for the equipment, and driver
cycle length constraints. A column generation algo-
rithm is developed for solving the static LP. The
model is tested by using the real data provided by
an LTL company. Networks with up to 40 nodes are
solved successfully. The effect of different driver route
length constraints is analyzed. The static model is
tested in a dynamic environment by optimizing the
network using the average demand, determining the
optimal trailer and tractor routes, then using those
routes for solving daily problems in which routes
are fixed but flows are allowed to vary. This strat-
egy results in solutions within an average 0.5% of the
optimum.
In the second part of this research, an opera-
tional version of the empty equipment allocation and
crew scheduling problem is considered. A dynamic
integer programming model is developed that opti-
mizes the detailed movements of trailers and trac-
tors given: demand constraints between consolidation
Transportation Science/Vol. 35, No. 1, February 2001 101

5. DISSERTATION ABSTRACTS
centers, conservation of flow constraints, driver cycle
length constraints and trailer route length constraints.
A solution method is proposed that involves column
generation to solve the LP relaxation, then branch-
and-bound to obtain an integer solution. The dynamic
model is tested by using the real data provided by the
LTL company, and networks with up to 30 nodes are
optimized successfully. The effects of different fore-
cast errors, tractor and trailer route length constraints
and period lengths are analyzed.
We conclude that both the static and dynamic mod-
els are computationally efficient for real problems,
and the combination of static and dynamic models
may be used for tactical and operational problems
faced by long-haul carriers. The static model may be
used for optimizing the routes for the relatively sta-
ble portion of the demand. On the other hand, the
dynamic model may be used for optimizing the routes
for the more fluctuating portion of the demand.
This thesis was submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy
in Industrial and Systems Engineering at the Univer-
sity of Southern California. The thesis supervisor was
Randolph Hall.
Integrated Control of Traffic Flow in
Corridor Networks
Christina Diakaki
Traffic congestion in corridor networks is notori-
ously increasing. According to field and simulation
investigations, a significant amelioration of traffic con-
ditions may be achieved as a result of the appropri-
ate integration of various control measures. It is the
scope of this thesis to develop a systematic, simple,
and generic methodology for this traffic control inte-
gration task.
The proposed IN-TUC strategy consists of three dis-
tinct interacting parts, the urban traffic control part,
the ramp metering part, and the route guidance part,
and produces traffic-responsive settings for the cor-
responding control actuators included in a corridor
network. For the ramp metering and the route guid-
ance parts, existing efficient methods are employed,
while a novel strategy has been developed for traffic-
responsive, network-wide urban traffic control. The
three parts of IN-TUC are integrated in the sense of
the mutual exchange of measurements and decisions.
This approach allows for the independent use of the
integrated control applications, and the easy extensi-
bility and expandability of the control system. More-
over, the approach ensures the functionality of the
system in case of failures or interruptions in any of
the integrated control applications. IN-TUC is built
upon well-known methods of the Automatic Control
Theory. This has led to efficiency, robustness with
respect to measurement inaccuracies, simplicity, and
transparency of the real-time code and of the design
procedure. Moreover, IN-TUC is generic enough to be
readily applicable (with only minor modifications) to
networks with arbitrary topology and characteristics.
Extensive simulation tests of the integrated strategy
and its parts indicated very high efficiency, which was
also demonstrated during the field-implementation
and evaluation of IN-TUC application to a part of the
Glasgow M8 corridor network in Scotland.
This thesis was submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in
Production Engineering at the Technical University of
Crete. The thesis supervisor was Markos Papageorgiou.
Econometric Models of Road Use, Accidents, and
Road Investment Decisions
Lasse Fridström
The present dissertation, which consists of three
essays in addition to an introductory overview, is an
attempt to understand and analyze road use, acci-
dents, and road infrastructure formation in a fairly
wide perspective.
In essay 1, “The barely revealed preference behind
road investment priorities,” the objective is to explain
the ranking of investment opportunities compiled
for the four-year National Road Plan 1990–93 by
the regional offices of the Norwegian Public Roads
Administration. The methodology used is a rank
order (exploded) logit modeling with sampling of
alternatives.
In essay 2, “Measuring the contribution of random-
ness, exposure, weather, and daylight to the variation
in road accident counts,” a methodology is developed
to assess the importance of certain factors that are nor-
mally not controllable by policy makers. The method
relies on generalized Poisson regression modeling.
In essay 3, “An econometric model of car owner-
ship, road use, accidents, and their severity,” the anal-
ysis of road accident counts is carried much further.
102 Transportation Science/Vol. 35, No. 1, February 2001

6. DISSERTATION ABSTRACTS
We estimate a recursive system of equations explain-
ing aggregate car ownership, overall and heavy vehi-
cle road use, rural and urban seat belt use, and injury
accident frequency and severity.
For input into the road use and accident frequency
equations, we develop an econometric method to esti-
mate exposure (overall and heavy vehicle kilometers)
from a combination of traffic counts, fuel sales statis-
tics, fuel prices, calendar data, weather conditions,
and vehicle pool characteristics.
Accident and casualty equations are estimated by
means of a specially designed procedure, called
IRPOSKML (Iterative Reweighted POisson-SKedastic
Maximum Likelihood), which allows the user to
unify a general nonlinear (Box-Cox) regression model
specification with the assumption that accident counts
follow the Poisson distribution.
This thesis was submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in
Transport Economics at the University of Oslo. The thesis
supervisors were Erik Biörn and Marc Gaudry.
Algorithms for Dynamic and Stochastic
Logistics Problems
Vijay S. Nori
In this thesis we study approaches for dynamic deci-
sion making under uncertainty in the context of two
important problems in logistics.
The Stochastic Inventory Routing Problem (SIRP)
addresses the coordination of inventory manage-
ment and transportation, and needs to be solved for
implementing a popular inventory strategy known as
vendor managed inventory replenishment. The goal
is to maximize the profit for the vendor by deter-
mining at any time the customers who should be
visited, the amount of product to be delivered, and
the sequence in which they should be visited. We
formulate the SIRP as a Markov decision process,
which is hard to solve optimally if a large number
of customers are involved. Approximation methods
are proposed which involve (i) developing an approx-
imate value function based on a decomposition of
the problem, (ii) developing an efficient randomized
approach to estimate an expected value, and (iii)
developing efficient ways to determine the action in
a state. Our research suggests that even though MDP
formulations are often deemed impractical for solv-
ing practical problems, with appropriate emphasis on
the computational aspects, fast and efficient methods
based on MDP models can be developed which solve
large instances of real-world problems.
The Online Resource Minimization Problem
(ORMP) is an optimization problem where work
with different deadlines arrives over time and has
to be completed before its deadline using valuable
resources. At a given set of decision points we decide
how much of the resource to purchase to complete
work in that period. The objective is to minimize the
total cost incurred, which is a function of the maxi-
mum amount of resource purchased in any period, to
finish all the work before their deadlines. We develop
several policies for the ORMP which decide on the
amount of resource to be purchased at any time based
on the amount of work remaining to be processed at
that time. We also present an optimal online policy
which is a very strong result seen rarely in literature
on online policies.
This thesis was submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in
Industrial and Systems Engineering at the Georgia Insti-
tute of Technology. The thesis supervisors were Anton
Kleywegt and Martin Savelsbergh.
Transportation Science/Vol. 35, No. 1, February 2001 103