Robust and dynamic models for supply chain and transportation networks

This dissertation considers operation and planning issues of dynamic supply chain and transportation networks in an uncertain environment. In particular, robust optimization approaches are applied to (1) emergency logistics planning, (2) network design and (3) congestion pricing problems under demand uncertainty residing in an appropriate uncertainty set such as box or polyhedral uncertainty set.;First of all, we develop a robust linear programming model of the cell transmission model (CTM) based on a robust optimization approach. Then, an affinely adjustable robust linear programming model is derived to study the multi-period problem. As an application area, we propose a methodology to generate a robust logistics plan that can mitigate demand uncertainty in humanitarian relief supply chains using a CTM based system optimum dynamic traffic assignment (SO DTA) model. Next, the proposed framework for SO DTA is extended to a dynamic network design problem. Finally, we consider robust congestion pricing problems under user equilibrium in static networks and extend it to consider robust dynamic user equilibrium optimal toll, which is formulated as a differential mathematical program with equilibrium constraints (DMPEC). Also, a cutting plane algorithm and a simulated annealing algorithm are proposed to solve the DMPEC problems.;Theoretically, the tractability and conservativeness of robust counterparts are discussed. Also, numerical experiments show that the robust optimization approach leads to high quality solutions compared to the deterministic problem or the sampling based stochastic problem. The results of the numerical experiments justify the modeling advantage of the robust optimization approach and provide useful managerial insights, which may have wider applicability in supply chain and transportation networks.