Design and Analysis of Novel Hybrid Multi-Objective Optimization Approach for Data-Driven Sustainable Delivery Systems

H. Giray Resat

doi:10.1109/access.2020.2994186

Abstract

This study presents a novel two-stage solution method designed for sustainable last-mile delivery systems in urban areas. A proposed hybrid solution methodology includes multi-criteria decision-making system to select the most efficient logistics providers by considering different performance indicators, and a mixed-integer linear optimization model for last-mile cargo distributions by drones within metropolitan areas. We present a multi-objective modeling approach by considering time windows for customer services and charging operations of drones and outline important characteristics of the mathematical programming problem to minimize transportation cost (in the meantime carbon dioxide emissions) and total sustainability score of the system by using epsilon constraint method to find out the Pareto frontiers. The main novelty of the proposed solution methodology is the inclusion of many performance indicators of last-mile delivery systems into multi-objective models for design of a sustainable city logistics. Additionally, the proposed model is applied to an illustrative case by using real-life data of one of the metropolitan in Turkey. The approach is shown as comparative analysis of proposed method with other two state-of-art solution methodologies for multi-objective problems, after defining some pre-processing, symmetry breaking steps, valid inequalities, and logic cuts.

Highlights

Companies operating in last-mile deliveries search for new and adaptive systems to survive under harsh competitive conditions
Traffic caused by urban cargo delivery accounts for about 10-15% of kilometers travelled in city centers and emits approximately 6% of all transport related greenhouse gas (GHG) emissions [1]. 20-25% of freight vehicle kilometers is related to goods leaving urban areas, and 40-50% is related to incoming goods
This study focuses on novel solution methodology for hybrid drone vehicle routing problem in an aspect of sustainability

Summary

Introduction

Companies operating in last-mile deliveries search for new and adaptive systems to survive under harsh competitive conditions. As a nature of last-mile delivery systems, it is the movement of cargoes from transportation hubs to final delivery destinations (mostly individual customers because of increased e-commerce transactions). Logistics centers are moved from the city centers to rural areas due to lower real estate costs, more multi-tier distribution systems are developed and expectation to adapt for the last-minute changes (especially on-demand services and urgent delivery requests) or maintenance/break-down cases leads to excess expenditures, longer delivery duration and increased rate of unsatisfied customers. Apart from the individual effect of last-mile delivery operations over citizens, there is a significant effect over energy consumption and environmental issues (such as greenhouse gas emissions, carbon footprint, etc.). Traffic caused by urban cargo delivery accounts for about 10-15% of kilometers travelled in city centers and emits approximately 6% of all transport related greenhouse gas (GHG) emissions [1]. The remaining percentage relates to internal exchange (i.e. goods having both their origin and destination within the city) [2]

Objectives

Methods

Results

Conclusion