Optimization Models for Last-Mile Service Operations

Abstract

Improving the quality of the last-mile service that comprises the movement of goods and people has been a recurrent theme in recent research. This dissertation aims to develop optimization models for addressing the emerging challenges encountered by three different last-mile service systems in the domains of inventory-delivery management, transportation and logistics, and emergency medical services, respectively.

The first chapter focuses on the domain of inventory-delivery management by investigating a two-echelon, single-product fulfillment system, in which the expected system-wide demand depends on the product’s price, the committed delivery time, and the number of local distribution centers in the system. The proposed model maximizes the expected profit while accounting for the total expected revenue, product holding costs, and fixed facility costs.

The second chapter concentrates on the domain of transportation and logistics, in which a fleet composition problem is studied. The proposed models consider using both internal truckload capacity and external less-than-truckload shipments for fulfilling stochastic demand. The objective is to minimize the total expected cost per period while determining the number of trucks to own for each truck type given a set of truck classes.

The third chapter contributes to the field of emergency medical services by studying a routing problem in the aftermath of a disaster, wherein the proposed models aim to determine the optimal routing strategy while maximizing the total expected number of survivors. A single ambulance bus is used to transport a given number of casualties to a medical center, accounting for time-dependent survival probabilities of the casualties.

The models and solution approaches developed in this dissertation provide managerial insights and can serve as a starting point when making decisions on supply chain design, facility locating, asset procurement, and fleet management and operations.