Dissertation
New ideas in delivery logistics: the consideration of travel time correlation and the use of package robots
University of Iowa
Doctor of Philosophy (PhD), University of Iowa
Autumn 2020
DOI: 10.17077/etd.005646
Abstract
This thesis seeks to explore several new ideas in delivery logistics. First, we examine a vehicle routing problem with a makespan objective incorporating both stochastic and correlated travel times. The presence of correlation between the travel time on arcs in a network is well acknowledged in the literature, but mostly omitted for the sake of simplicity. In our work, we utilize ideas from extreme-value theory to build high quality estimates of the distribution of the makespan for a set of routes. We then discuss how our approach can be embedded within any vehicle routing heuristic to find the minimumexpected makespan. Finally, we demonstrate the impact of including correlation in a routing context through a number of numerical tests. We find correlation is especially important to consider when there is a high correlation for arcs near a particular incident.
Second, we explore the potential of ground-based robots in the context of parcel deliveries. We begin by investigating a two-tier delivery network with robots operating on the second tier. We determine the optimal number of local robot hubs as well as the optimal number of robots to service all customers and compare the resulting operational cost to conventional truck-based deliveries. We solve the problem using mixed integer programs and show that robots can save up to 90% of operational costs in comparison to conventional truck based deliveries. Finally, we extend this problem by considering pedestrian level of service (LOS) zones with stochastic travel times and soft time windows. Pedestrian LOS is a measure of pedestrian flow density, and is an important characteristic for sidewalk robot-based deliveries. Our model considers service costs in terms of early and late arrivals, and we propose a path selection process given the existence of pedestrian LOS zones. We conclude that robots travel different paths to deliver packages depending on the pedestrian LOS in zones.
Details
- Title: Subtitle
- New ideas in delivery logistics: the consideration of travel time correlation and the use of package robots
- Creators
- Iurii Bakach
- Contributors
- Ann Campbell (Advisor)Jan Ehmke (Advisor)Ray de Matta (Committee Member)Weimin Han (Committee Member)David Stewart (Committee Member)
- Resource Type
- Dissertation
- Degree Awarded
- Doctor of Philosophy (PhD), University of Iowa
- Degree in
- Applied Mathematical and Computational Sciences
- Date degree season
- Autumn 2020
- DOI
- 10.17077/etd.005646
- Publisher
- University of Iowa
- Number of pages
- xii, 141 pages
- Copyright
- Copyright 2020 Iurii Bakach
- Language
- English
- Description illustrations
- color illustrations
- Description bibliographic
- Includes bibliographical references (pages 89-100).
- Public Abstract (ETD)
The importance of last-mile delivery, the final step in the retail supply chain, cannot be overestimated. With the recent explosion of e-commerce, logistics service providers are exploring alternative approaches to keep customers happy. Some of these approaches are quite innovative and require a closer look in order to evaluate their potential. In this dissertation, we look at one solution, namely, ground-based package robots. Specifically, we aim to explore how they can be used and what the benefits of their utilization are. We investigate a two-tier delivery network with robots operating on the second tier. Then, we explore the effect of pedestrian traffic on robot effectiveness and customers satisfaction level. In addition, we explore the effect of the correlation between the travel time on arcs in a network on routes. We demonstrate that the impact of the correlation can be significant and highlight the situations when it can have a big impact on routing solutions.
- Academic Unit
- Interdisciplinary Graduate Program in Applied Mathematical & Computational Sciences
- Record Identifier
- 9984036791402771
Metrics
10 File views/ downloads
74 Record Views