I am an assistant professor in Industrial Engineering at Georgia Tech. I am interested in the theory and applications of supply chain and logistics models primarily involving transportation and inventory decisions, and in related optimization topics. The methodologies I use to study these problems include linear, mixed-integer and dynamic programming.
My CV. (Last updated November 6, 2014.)
Submitted or Under Review
We propose dynamic linear programming games to model situations where uncertain costs must be shared over time by agents whose attitudes towards uncertainty are captured by coherent dynamic risk measures. Whereas cooperation is always possible when the agents are risk neutral, it proves much more difficult in the presence of risk aversion.
We study the natural extensions of blocking and anti-blocking polyhedra to arbitrary dimensions and give sufficient conditions for complementary slackness and integrality of extreme points.
We show that an approximate linear programming (ALP) bound for the TSP, introduced in earlier work, is equivalent to the well-known Held-Karp bound. We then show how the ALP bound can be extended to several TSP variants.
We propose an infinite linear programming model for long-term workforce planning of a single worker type, e.g. nurses, in a large health care system. We give a series of conditions a system of this kind should satisfy, and use them to prove the optimality of a natural lookahead policy.
Honorable Mention, INFORMS SPPSN Best Paper Award, 2013.
We explain how Optimized Financial Systems uses customers' financial information and actuarial data in an optimization model to give recommendations on investments and tax planning, particularly focused on individual retirement arrangements (IRAs).
Published or Accepted
We assess the current U.S. pediatric nurse practitioner workforce and investigate the impact of potential policy changes to address forecasted shortages.
We propose a TSP with stochastic arc costs in which the salesman can observe outgoing arc costs at every city before choosing where to go next. We use approximate linear programming to tractably bound the problem and construct high-quality policies.
We study an agricultural supply chain with stochastic demand for perishable goods, where several suppliers can save on transportation costs by consolidating shipments, and propose a simple look-ahead heuristic.
We propose a framework of lower bounds for the asymmetric TSP based on approximating the dynamic programming formulation. We then introduce an exact reformulation that generates a family of polynomially-solvable, successively tighter lower bounds. We show that the base member of this family yields a bound greater than or equal to the well-known Held-Karp bound.
For lot sizing situations in which multiple retailers or producers can pool orders, we construct allocations dynamically over time so that costs are met as they are incurred and no one has the incentive to defect at any point.
We consider cooperative traveling salesman games with non-negative asymmetric costs satisfying the triangle inequality. Using a variant of the Held-Karp relaxation and its dual, we construct a stable cost allocation with budget balance guarantee equal to the Held-Karp integrality gap for the asymmetric traveling salesman problem.
We evaluate the California cut flower industry’s current transportation practices and investigate the feasibility and cost of establishing a shipping consolidation center in Oxnard, California.
We introduce a fixed-charge transportation problem with linear product blending and give a polyhedral analysis. Applications include transportation problems in the petrochemical, energy and agriculture industries.
We study models for piecewise linear data fitting (regression), including new mixed-binary models with variable regions. Applications include approximate inventory valuation.
We characterize the value function of a discounted, infinite-horizon variant of the single-item production lot-sizing problem and show that it inherits several structural properties from finite mixed-integer program value functions.
We introduce a time decomposition for inventory routing problems that depends on approximately valuing inventories at suppliers and consumers. Computational experiments use maritime inventory routing instances.
We develop an algorithm to calculate how stable a solution to a binary mixed-integer program is with respect to cost changes. Applications include real-time decision making scenarios, such as iterative combinatorial auctions.
* Indicates supervised student co-author.
International Conference on Logistics and Maritime Systems (LOGMS). Rotterdam, the Netherlands. August, 2014.
Operations/Management Science Workshop. Booth School of Business, University of Chicago. Chicago, Illinois. December, 2013.
Department of Industrial and Operations Engineering, University of Michigan. Ann Arbor, Michigan. November, 2013.
INFORMS Annual Meeting. Minneapolis, Minnesota. October, 2013.
DOS Seminar Series. Stewart School of Industrial and Systems Engineering, Georgia Tech. Atlanta, Georgia. September, 2013.
Tutte Seminar Series. Department of Combinatorics and Optimization, University of Waterloo. Waterloo, Ontario. May, 2013.
VII Latin American Algorithms, Graphs and Optimization Symposium. Playa del Carmen, Mexico. April, 2013
ISyE 3103 - Supply Chain Modeling: Logistics
ISyE 3133 - Engineering Optimization
ISyE 4301 - Supply Chain Economics
ISyE 7203 - Logistics Systems Engineering
ISE 330 - Introduction to Operations Research: Deterministic Models (USC)
ISE 532 - Network Flows (USC)
Shabbir Ahmed, Christiane Barz, Maged Dessouky, Alan Erera, Ricardo Fukasawa, Will Haskell, Fatma Kılınç-Karzan, Mariel Lavieri, Jim Moore, George Nemhauser, Christine Nguyen, Dimitri Papageorgiou, Luis Rademacher, Martin Savelsbergh, Nelson Uhan, Juan Pablo Vielma, Joshua Woodruff.