Parametric Design of Multi-Echelon Last-Mile Delivery Systems Using Novel Technologies

Abstract

The growing demand for fast and cost-effective last-mile delivery has prompted logistics operators to explore emerging delivery technologies such as autonomous vehicles (AVs) and drones. This dissertation investigates the design and strategic planning of multi-echelon last-mile delivery systems using emerging delivery technologies. Through three complementary studies, we explore the integration of AVs and drones into traditional truck-based delivery networks using analytical models and parametric design frameworks.

The first study examines recipient-dependent routing policies in AV-enabled deliveries, where recipients participate by using their AVs for parcel pickup. We introduce and analyze three delivery policies- AV pickup, Hybrid pickup, and Crowdsourced delivery— and derive upper and lower bounds on their total system costs. These policies are compared against the traditional truck delivery model, with results showing that recipient-dependent strategies become increasingly dominant as the number of deliveries grows.

The second study extends this concept to the delivery of perishable goods, where delivery time is critical. Using parametric design, we introduce length-constrained (time-sensitive) routing models and compare AV, Hybrid, and crowdsourcing policies to traditional truck delivery. Our findings show that AV integration offers significant advantages in reducing delivery failure for perishable goods, particularly under increasing hand-off costs and high delivery density.

The third study focuses on tactical fleet planning and routing in drone-assisted last-mile delivery. We model a mixed fleet of trucks and drones, considering drone range, truck and drone capacities, multi-launching, and vehicle synchronization. Using continuous approximation, we derive closed-form solutions for fleet composition and routing strategies under varying operational scenarios. The analysis yields policy spaces that guide optimal system design based on drone and truck operational parameters, such as vehicle speeds and capacity limits.

This dissertation, through its integrated studies, offers a unified framework for the strategic design and operational evaluation of multi-echelon last-mile delivery systems using emerging delivery technologies. The models developed provide managerial insights for logistics planners, policymakers, and researchers seeking scalable and efficient delivery solutions.