A Comprehensive Evaluation of Weak and Strong Mutation Mechanisms in Evolutionary Algorithms for Truck Scheduling at Cross-Docking Terminals

Abstract

Freight transportation is important for the national economy in many countries. An efficient distribution of products within supply chains may lower the associated costs and allow setting competitive prices to increase the number of sales. Many supply chain players use the cross-docking terminals to facilitate the cargo distribution process. An effective scheduling of the arriving trucks at the cross-docking terminals is critical to ensure their timely service. A number of Evolutionary Algorithms have been developed to solve the truck scheduling problem, some of which apply strong mutation for altering solutions throughout the search process, while the rest rely on weak mutation without providing any justification for applying a specific mutation mechanism. This study performs a comprehensive comparative analysis of the strong and weak mutation mechanisms. Furthermore, a novel heuristic algorithm, which accounts for the truck service priority and the truck service order restrictions, is proposed for initializing the chromosomes and population. The truck scheduling problem at a cross-docking terminal is formulated as a mixed integer programming model, minimizing the total weighted truck service cost. An Evolutionary Algorithm is designed to solve the problem. Two categories of the Evolutionary Algorithm, one of which applies strong mutation, while the other one relies on weak mutation, are evaluated based on various performance indicators. Results demonstrate that deployment of weak mutation improves the objective function value at termination on average by 10.8% as compared with strong mutation without affecting the computational time substantially. The analysis also shows that weak mutation yields more diverse population. Moreover, the proposed heuristic for initializing the chromosomes and population outperforms the initialization mechanisms that are commonly used in the literature.