Multi-Objective Whale Optimization Algorithm And Applications In Vehicle Routing Problem With Random Demands

In recent years,new generation information technologies represented by artificial intelligence have become an important technical guarantee and core driving force for China’s social innovation and development.Swarm intelligence optimization algorithms,as a popular branch of artificial intelligence research,have been widely valued and applied in various engineering fields.Whale Optimization Algorithm(WOA)is a new type of swarm intelligence optimization algorithm proposed in recent years.WOA has the advantages of simple structure,few setting operators,fast convergence speed,and good balance between local and global search,and has been widely applied in optimization problems,machine learning,data mining,and other fields.In addition,with the continuous development of e-commerce,solving the "last mile" problem in community logistics has become a focus and difficulty in current research.To solve this problem,it is necessary to minimize transportation time,driving distance,and other objectives while meeting residents’ needs.Such problems can be described as multi-objective vehicle routing problems with stochastic demands(VRPSD).Since customer demands in VRPSD are unknown before the vehicle arrives,it is difficult to find a perfect solution.However,the development of swarm intelligence algorithms makes it more efficient and accurate to solve vehicle routing problems(VRP).Therefore,applying swarm intelligence algorithms to VRPSD has good practical significance and technical support.This study aims to provide a new idea for solving multi-objective problems by applying Whale Optimization Algorithm(WOA)to VRPSD.The specific content arrangement is as follows:1)The biggest advantage of WOA lies in its excellent global search capability,but it also has some shortcomings,such as the possibility of getting stuck in local optima in the later stages of the algorithm,and slow convergence speed when dealing with multi-objective problems.Therefore,this study first needs to address the inherent limitations of the algorithm,and then further improve the algorithm to solve multi-objective problems.Specifically,we increase the population diversity through the quasi-reflective learning mechanism and Logistic chaotic mapping.Secondly,we introduce the FADs vortex effect and wavelet mutation from the Marine Predator Algorithm(MPA)in the search stage to enhance the stability and ability to escape local optima in the early and later stages of the algorithm.Finally,a nonlinear segmented convergence factor is proposed to balance the algorithm’s local and global search capabilities.Experimental results show that the improved algorithm has better stability and ability to escape local optima,and the convergence speed and optimization accuracy are also improved.2)The difficulty of solving multi-objective VRPSD problems lies in the conflicting optimization goals and the need to consider the priorities of the objectives,the complexity of computation,and the choice of measurement standards.To address these issues,this study analyzes the characteristics of multi-objective problems and the shortcomings of WOA in handling discrete problems.Based on the improved algorithm,the study introduces the concept of cooperative evolution,an adaptive dynamic layering mechanism,and a fast non-dominated sorting strategy to improve the precision of the algorithm in solving multi-objective problems while reducing the complexity of computation,while ensuring the convergence speed and global performance of the algorithm.Finally,the improved algorithm is applied to multiple multi-objective VRPSD problems.The experimental results show that the improved WOA can effectively reduce the cost and vehicle travel distance,improve delivery efficiency,and have higher practicality and application prospects.