Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Business

Program Name/Specialization

Operations and Supply Chain Management

Faculty/School

Lazaridis School of Business and Economics

First Advisor

Salar Ghamat

Advisor Role

Advisor

Second Advisor

Borzou Rostami

Advisor Role

Advisor

Abstract

This dissertation investigates how technology can improve operational decisions and resource allocation in service and production systems with limited capacity. I develop mathematical models and solution methods for settings in which congestion, heterogeneous demand, and technology-specific limitations interact with strategic and operational decisions. Across these settings, I examine how technology changes the way demand is served, workload is distributed, and constrained resources are utilized. Using analytical, optimization, and learning-based methods, I show that adopting technology does not necessarily reduce costs or alleviate capacity constraints. Its value depends on how organizations design incentives, allocate demand, manage congestion, and account for its effects on workload and reliability.

In the first essay, I study how remote monitoring can improve access to maternal healthcare when clinical capacity is limited. Patients with high-risk pregnancies may require frequent monitoring, but repeated in-person visits impose travel and waiting costs on patients and increase the provider’s workload. I develop a model in which a payer determines reimbursement for virtual care, a provider assigns patients to remote or in-person monitoring based on their risk levels, and patients value the two care options differently in terms of cost and convenience. I also examine how the provider determines the frequency of virtual visits and how reimbursement and capacity constraints influence this decision. I find that targeted reimbursement and flexibility in setting the frequency of virtual visits can improve coordination and support greater use of remote monitoring.

In the second essay, I study the assignment of different types of jobs to a network of facilities where limited capacity creates congestion and waiting. I consider facilities that process jobs either in their order of arrival or according to assigned priorities. I develop exact mixed-integer second-order cone programming models that account for assignment costs and waiting times. I then develop machine-learning-augmented optimization models that approximate waiting-time components and can be applied to larger problems. Computational results show that these models provide highly accurate solutions while substantially reducing computational time relative to the exact models.

In the third essay, I study how a firm can combine conventional manufacturing and additive manufacturing, commonly known as 3D printing, to support spare-parts logistics. The firm must determine which units of each part should be supported through conventionally manufactured stock and which should be printed on demand. These decisions are affected by differences in the reliability of conventionally manufactured and printed parts, as well as by congestion created when multiple parts share limited printer capacity. I develop a long-run cost model and an exact optimization approach that jointly account for inventory costs, printing costs, source-dependent reliability, and waiting costs. I find that combining conventional manufacturing and additive manufacturing can be optimal and that printed-part reliability and shared printer congestion can substantially affect the preferred support strategy.

Convocation Year

2027

Convocation Season

Spring

Download

Available for download on Saturday, September 29, 2029

Share

COinS