Addressing Challenges in Public Service Operations Management: Data-Driven Solutions and Strategies

This dissertation examines data-driven decision-making in crucial areas of public service operations management, with a specific focus on liver allocation and child welfare operations. 

Chapter 1 provides an overview of the research background and describes the common challenges in public service resource allocation as well as context-specific operational complexities.

 Chapter 2 introduces a decision support model for split liver transplantation (SLT) to enhance efficiency and fairness in liver allocation. Through a multi-queue fluid system model, optimal matching procedures are identified, demonstrating the potential benefits of increased SLT utilization. 

Chapter 3 explores learning-informed algorithms for SLT resource allocation, utilizing a multi-armed bandit (MAB) model to balance exploration and exploitation in surgical team selection. Novel algorithms, L-UCB and FL-UCB, are developed and shown to exhibit superior performance in allocating organs while incorporating experience-based learning and fairness concerns.

 Chapter 4 studies the impact of workload on the screening of child maltreatment reports, highlighting the need for load-aware risk protocols in human-AI collaborations. 

Chapter 5 concludes the dissertation by outlining future research avenues and potential operational improvements in public service.