An Improved Tabu Search Algorithm for the Stochastic Vehicle Routing Problem With Soft Time Windows

Guoming Li,Junhua Li

doi:10.1109/access.2020.3020093

Guoming Li, Junhua Li

Open Access

PDF Available

https://doi.org/10.1109/access.2020.3020093

Copy DOI

Export

Save

Cite

Journal: IEEE Access	Publication Date: Jan 1, 2020
Citations: 29	License type: CC BY 4.0

Affiliation: Nanchang Hangkong University

Abstract
Highlights/Summary
Full-Text PDF
Similar Papers

Abstract

Listen

In real-world environments, vehicle travel and service time will be affected by unpredictable factors and present a random state. Because of this situation, this article proposes the vehicle routing problem with soft time windows and stochastic travel and service time (SVRP-STW). The probability distribution of vehicle travel and service time are introduced into the model, and a stochastic programming model with modification is established to minimize the distribution cost. An Improved Tabu Search algorithm (I-TS) based on greedy algorithm is proposed, in which adaptive tabu length and neighborhood structure are introduced; the greedy algorithm is used instead of the random methods to generate the initial solution. Experiments on different scale instances prove the effectiveness and superiority of the proposed algorithm.

Highlights

With the rapid development of the distribution industry, the expanding scale of logistics distribution, and the enhancing complexity of urban transportation networks, some phenomena occur frequently, e.g., unreasonable distribution route planning, empty distribution vehicles, and unreasonable order allocation
A new stochastic programming model is proposed considering the stochastic vehicle travel and service time, and an improved meta-heuristic algorithm is proposed considering the high complexity of the SVRPTW problem
In the process of applying the improved tabu search algorithm, the variance of vehicle travel time and service time was set as 10%, 20%, 30% and 40% of the mean respectively, and a comparative analysis was carried out

Summary

INTRODUCTION

With the rapid development of the distribution industry, the expanding scale of logistics distribution, and the enhancing complexity of urban transportation networks, some phenomena occur frequently, e.g., unreasonable distribution route planning, empty distribution vehicles, and unreasonable order allocation. It has been used to solve the VRP and its variants, e.g., VRP [11], VRPMTW [12], HFVRP [13], ConVRP [14] and MCVRP [15] In this context, a new stochastic programming model is proposed considering the stochastic vehicle travel and service time, and an improved meta-heuristic algorithm is proposed considering the high complexity of the SVRPTW problem. They analyze that the arrival time distribution would be truncated at the earliest time windows, affecting the estimation of the arrival time of subsequent customers To solve this problem, a specific statistical method is proposed to obtain the cumulative probability distribution of the vehicles over the customers, and a meta-heuristic algorithm based on Iterative Local Search (ILS) is proposed.

ANALYSIS OF THE OBJECTIVE FUNCTION

EXPERIMENTS AND RESULTS

Findings

CONCLUDING REMARKS AND FURTHER RESEARCH