Dynamic Games and Robust Models for Transportation and Service Networks

Open Access
Zhang, Bo
Graduate Program:
Industrial Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
April 25, 2013
Committee Members:
  • Tao Yao, Terry L Friesz, Dissertation Advisor
  • Tao Yao, Terry L Friesz, Committee Chair
  • Paul Griffin, Committee Member
  • Susan H Xu, Special Member
  • Urban Freight Transportation
  • Truckload Service Procurement
  • Dynamic Stackelberg Game
  • Two-Stage Robust Optimization
This dissertation reviews the theories and methods of dynamic games in traffic assignment and optimization under uncertainty, refines and applies them to address the following two specific research problems: (1) Urban freight transportation planning. Urban freight transportation supports the economic and social development of an urban area. However, there are negative externalities associated with urban freight transportation such as emission, noise and congestion. We consider that there exists a metropolitan planning organization (MPO) that is responsible for reducing the congestion caused by freight transportation on the urban road network. In this study, based on different assumptions on the MPO’s capability to influence freight truck traffic, we develop two dynamic game-theoretic models to support the MPO’s goal of reducing congestion caused by freight transportation. Heuristic algorithms are developed to solve the proposed models. We conduct numerical analyses to derive optimal urban freight transportation plans. Our analyses show that the congestion caused by urban freight transportation can be effectively reduced by properly controlling the freight truck traffic. This can provide insights to the MPO into urban freight traffic regulation. (2) Truckload service procurement under uncertainty. Truckload (TL) transportation is a necessary component of a shipper’s logistics system and the associated procurement expenditure accounts for a significant portion of the shipper’s overall cost. In this study, we argue that a shipper’s TL service procurement cost can be reduced by appropriately handling the uncertainty during the procurement process. In particular, based on different assumptions on the availability of distributional information of the uncertain shipping demand, we propose two robust models to reduce the shipper’s cost. We develop new solution approaches and conduct numerical tests on real-world sized instances of TL service procurement to demonstrate the applicability of the proposed robust models. Our analyses show that the solutions to robust models yield a lower procurement cost than the solution to a deterministic model. Insights into the design and operation of TL service procurement are drawn from the numerical analyses.