Gainesville, FL 32611
Peter Yun Zhang
Ph.D. Candidate, Massachusetts Institute of Technology (MIT)
Title: Dynamic Robust Optimization for Inventory Network Design
Abstract: I define a class of network design problems that are motivated by modern operations research applications, from public health networkdesign to e-commerce platform coordination. I formulate the decision problem as a dynamic optimization problem on networks. Such multistage problems are intractable in general. I characterize a class of heuristics that are suitable for solving such problems, trading off tractability and optimality. More precisely, I show that the simplest of such heuristics (affine policies) are optimal for simple networks (trees). Illustrating the decision framework with an important application, Ifocus on the design of an antibiotic supply chain to defend against
Bioattacks, i.e., the intentional release of pathogens or biotoxins against humans to cause serious illness and death, pose a significant threat to public health and safety due to the availability of pathogens worldwide, scale of impact, and short treatment time window. I model the defender’s static antibiotic inventory prepositioning decision and dispensing capacity installation decision, attacker’s move, and defender’s adjustable shipment decision, so as to minimize inventory and life loss costs, subject to population survivability targets. I explicitly account for the strategic interaction between defender’s and attacker’s actions, assuming information transparency. I perform a high-fidelity case study on the design of an antibiotic supply chain with millions of nodes to guard against anthrax attacks. I calibrate the model using data from a wide variety of publicly available sources, including literature and field experiments, and produce policy insights regarding inventory prepositioning and demand prediction.
Department of Industrial and Systems Engineering at the University of Florida