Research On Intelligent Algorithm For Vehicle Routing Problem

Vehicle routing problem occupies an important position in the field of logistics management.It has strong practicability and becomes a hot topic in the field of operations research and management science.However,the high complexity of the problem makes it difficult to achieve high efficiency and low cost distribution,overcome these difficulties is the core problem that puzzles manufacturers and scholars.At present,the research of vehicle routing problem is mainly focused on the design of the intelligent algorithm with high search capability.Capacitated vehicle routing problem(CVRP)in the needs of each customer demand can only be delivered by one vehicle.When it is not constrained by this condition,it becomes a Split Delivery Vehicle Routing Problem(SDVRP).This paper: first of all,summarizes the CVRP's and SDVRP's research and development,the mathematical model,the characteristics of the solution and so on;Secondly,aiming at CVRP,a the coding method for particle swarm algorithm and the adjustment strategy for infeasible solution are designed,four improved particle swarm optimizations which Separately composed of two kinds of location update formulas(random perturbation and Gauss perturbation)and two local search strategies(neighborhood search and single point mutation)are improved.Some experimental results show that the algorithm GBPSO(Gauss perturbation with single point mutation)has better search ability and higher search efficiency.Thirdly,for SDVRP,a kind of clustering algorithm,which arranges paths after grouping is proposed.Considering the balance of vehicle load and characteristics of feasible solution and combining with the distance between customers and load,set a split threshold and a reasonable split strategy.While all the customers are divided into groups,the customers paths for each vehicle is arranged by Ant Colony Optimization algorithm.The experiment results show that the proposed algorithm has higher stability and better results in SDVRP.Finally,it summarizes the work of this paper and points out that the field needs to depth study the diversity of logistics distribution model and improve the search ability of the algorithm in future.