Ant Colony Algorithm For Vrp Problem With Time Windows

As the intensify of market increasingly and the enhance of global ecnomics,logistics became an important measure for enterpnses to promote the power of market and core competition.Thereinto Vehicle Routing Problem(VRP) of logistics distribution influence on service quality,logistics cost and benifit of enterprise,so it's an important problem to solve the problem.In particular,as prevalence of Internet and development of electronic commerce,It is impossible to use traditional VRP arithmetic to deal with quick response of logistic divition for customer,upon that the concept ot time windows emerge as the times require.Vehicle Routing Problem with Time Windows(VRPTW) is an NP-hard complete optimization problem,which more complex than Vehicle Routing Problem.The solution method was divided to precison algorithm and heuristic algorithm.Precison algorithm can't keep away from exponent blast problem be acause of it's strict mathematics method and can solve brushfire VRPTW only.In virtue of VRPTW is a NP-complete problem, we can find approximation algorithm only.Therefor most people spent time to conformation of heuristic algorithm with high quality.Among these heuristic algorithm,Ant Colony Optimization algorithm became more and more popular because of it's robustness,distributed computation and the excellency of easier to combine with other algorithm tools.However ACO have the antinomy between search space and performance,easily to convergence to solution which is not globally optimal solution and chink of long calculate time.In addition,the choice of ACO parameter by right of experience u5ually.Aim at these problems,this dissertation develops the rearch based on the predecessor works.The main works are as follows:1.Research on the choice of ACO parameter by a series of emulation experiment, bring the optimal ACO parameter combination.Compare with the of parameter by experience,ACO parameter combination improved the efficiency of algorithm.2.Apply ACO algorithm to solution of VRPTW.An approaoh to improve translation rule and pheromone updating rule was applied by introducing the comcept ot EVenness of Solution,Selection Window and Ant Attracion,which conformation of a dynamic ACO algorithm.In this approach,ant will adjustable for different situation.Simulation results show that the dvnamic ACO algorithm can settle the contradictory between convergence speed and stagnation behavior efficiently and is very suitable for solving VRP.3.Base on analyse the multiple operation of logistics division,research and design a logistics division system which can both deal with handcraft and computer aimed at mutiple operation demand.