Back to search

McMaster University

Optimal Spatial and Temporal Decisions under Uncertainty

Abstract

dc:description.abstract

This dissertation studies two interconnected problems in operations management: the design and deployment of autonomous indoor delivery systems in multi-floor buildings, and the strategic behavior of customers competing for limited inventory in high-demand sales events. This dissertation is organized into five chapters. Chapter 1 introduces the motivating challenges and outlines the problems this work sets out to address. Chapter 2 studies the optimal location of a central hub for a fleet of autonomous delivery robots operating in a multi-floor building. The hub serves as the home base to which robots return between service requests, and its location jointly determines horizontal travel distance, vertical inter-floor travel cost, and queuing delay. We show that the optimal horizontal position on any given floor is the Scaled Gravity Center, the floor-demand-weighted spatial mean of the demand distribution, a closed-form result whose approximation quality is validated statistically through a Wilcoxon signed-rank test. Embedding an M/M/1 elevator queuing model into the spatial cost structure yields a unified performance functional that captures horizontal travel, vertical displacement, and stochastic waiting time jointly. We characterize the optimal hub floor through a discrete convexity condition, establishing parametric bounds under which a local sign test on first-order differences of the cost function is sufficient for global optimality. Numerical experiments comparing the integrated model against two benchmarks, the Highest-Peak heuristic, which places the hub on the peak-demand floor, and the Distance-Only policy, which ignores queuing delay, show that the integrated model yields cost savings of 14 % to 25 % over the Highest-Peak heuristic across all asymmetric demand profiles considered, and that ignoring queuing delay leads to additional losses when demand is unevenly distributed across floors. Chapter 3 examines the fleet architecture decision facing an autonomous delivery system operator: whether to dedicate robots to individual floors or to share a combined fleet across all floors. We derive closed-form expressions for expected response time under both architectures and characterize conditions under which each is preferable. The main result is the ``Sandwich Policy'': when the vertical travel penalty between floors is below a critical threshold, there exist two demand thresholds such that non-pooling is optimal at low demand, Pooling is optimal at intermediate demand, and non-pooling is again optimal at high demand. We further derive the critical vertical penalty above which no pooling dominance interval exists, and extend the two-floor results to buildings with an arbitrary number of floors. Chapter 4 develops a unified analytical framework for customer arrival timing in high-demand sales events where a limited number of units is allocated first-come, first-served to a population of competing customers. On the non-strategic side, we show that the optimal arrival time across all competitor configurations is governed by a single hazard-rate condition, derive a closed-form arrival rule under uniformly distributed competitors, and identify a critical scarcity threshold below which rational customers find it optimal to queue overnight regardless of waiting cost, providing the analytical explanation for multi-day queues at events such as the Taylor Swift Eras Tour. On the strategic side, we formulate a two-player effort game, establish the existence and uniqueness of Nash equilibrium, derive the closed-form equilibrium, prove the comparative statics analytically, and compute the price of anarchy in closed form, showing that all competitive effort is socially wasteful and the welfare loss fraction equals the ratio of waiting cost to product value. Chapter 5 summarizes the contributions of the dissertation and outlines directions for future research.

Degree

thesis:*
Department dc:contributor.department
Business
Year dc:date.issued
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Ulubayova, Fatima
Advisors dc:contributor.advisor
  • Zhou, Yun
  • Parlar, Mahmut

Rights

Language dc:language.iso
en_US

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
McMaster University
Base URL
macsphere.mcmaster.ca/server/oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
related terms
citation

Ulubayova, Fatima. Optimal Spatial and Temporal Decisions under Uncertainty. 2026. https://hdl.handle.net/11375/33749