Graduation Year

2026

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Industrial and Management Systems Engineering

Major Professor

José L. Zayas-Castro, Ph.D.

Committee Member

Jorge Acuña, Ph.D.

Committee Member

Hadi Gard, Ph.D.

Committee Member

Robert Frisina, Ph.D.

Committee Member

Jay Wolfson, Dr.P.H.

Keywords

Emergency medical services, Equity and efficiency, Game theory, Kidney transplantation, Reinforcement Learning, Stochastic programming

Abstract

The allocation of scarce medical resources presents one of the most consequential challenges in health care, requiring decisions that balance equity and efficiency under pervasive uncertainty. This dissertation develops optimization-based frameworks for two critical domains, organ transplantation and emergency medical services, and demonstrates how operations research methods can improve upon current practice along both dimensions simultaneously.

In the domain of deceased-donor kidney allocation, two complementary frameworks are proposed. The first is a bi-objective stochastic optimization model that jointly maximizes post-transplant survival (efficiency) and prioritizes candidates with the highest pre-transplant mortality risk (equity), subject to regional chance constraints on graft failure rates. The Nash bargaining solution identifies a fair operating point on the Pareto frontier, and the resulting weights are deployed in an online adaptive two-stage stochastic program that makes allocation decisions as organs become available. Applied to national OPTN data, the framework increases monthly deceased-donor transplants by 4.5%, reduces average graft failure probability by 6.22%, improves average expected survival among recipients by 22.95%, and expands access for the highest-urgency candidates by 15% relative to actual OPTN allocations. The second framework extends this work to the individual patient–organ level through a multiperiod two-stage stochastic program that integrates pre-transplant quality-adjusted life years to quantify medical urgency alongside post-transplant QALYs to measure expected benefit. A four-state Markov chain Monte Carlo simulation generates clinically grounded QALY scenarios, and the model incorporates blood-type-identical matching bonuses, top-20% longevity matching, waiting time prioritization, and age-group proportionality bounds consistent with OPTN's continuous distribution framework. Applied to OPTN Region 5, this model increases transplant volume by 23.8%, primarily by utilizing higher-KDPI organs that the current system discards, while raising blood-type-identical matching from 92.1% to 99.7% and nearly doubling the number of top-tier longevity-matched pairs.

In the domain of emergency medical services, this dissertation introduces an integrated optimization-to-learning framework for real-time ambulance-to-ED allocation. An enhanced progressive hedging algorithm solves a multi-stage stochastic program across an expanded scenario tree, generating near-optimal routing decisions under demand uncertainty. A reinforcement learning agent trained on these solutions serves as a computationally efficient surrogate, achieving statistically equivalent performance to the stochastic program — within 0.1% of its objective value — while reducing inference time to approximately one millisecond per decision, enabling real-time deployment in live EMS dispatch systems.

Finally, the dissertation presents a policy-oriented discussion that translates these quantitative findings into actionable recommendations for the transplant community and policymakers, supported by economic analysis demonstrating that optimization-based allocation of organs already recovered could generate hundreds of millions of dollars in annual Medicare savings. Collectively, this work demonstrates that coordinated, mathematically grounded allocation frameworks can simultaneously improve system performance and patient-level fairness across multiple health care settings.

Download

Share

COinS