Rearch On VRPSD Based On Particle Swarm-ant Colony Hybrid Algorithm

With the rapid development of economic globalization and information technology, Logistics has risen worldwidly as an important service industry.However,with the increase in the frequency of economic activity and customer consumption levels,there will be a lot of uncertain information,so how to use the information to create wealth become a key question.The study of vehicle routing problem,especially for vehicle routing problem with stochastic demand,can help the logistics companies arrange the dilivery routes reasonably,save the transportation costs,improve the service levels and customer satisfaction.So as to get the wealth of"the third profit source".All of these are of great significance.Under a systematic analysis on the related research achievements and the main problems of the vehicle routing problem at home and abroad,this paper was aimed to solve a series of problems,such as large-scale vehicle routing problem was difficult to solve,the solution precision was not high and there were many shortcomings on using a single algorithm to solve the model.Connected with the actual situation of transport in logistics,this paper was carried out a deep investigation to large-scale vehicle routing problem with stochastic demand.Firstly,due to the characteristics of vehicle routing problem with stochastic demand,vehicle routing problem with stochastic demand model which was in the state of transportation (pick up) was built in the study.In the modle the count of failure route was setted to twice and the strategy of part-service was allowed.Secondly,for the large-scale vehicle routing problem difficult to solve,the two-phase heuristic algorithm was applied to solve the model.In the first phase, K-means cluster analysis algorithm was used to divide the large-scale customer network into several small-scale VRP,in order to reduce the computation and improve the calculation speed and accuracy.In the second phase,particle swarm optimization algorithm was combined with ant colony optimization algorithm to form the particle swarm-ant colony hybrid algorithm.So that the modle-solving process was simplified,and the vehicle routing problem with stochastic demand was solved quickly and effectively.Finally,an automatic optimization on routes was effectively worked out with the MATLAB progranmming.An example was used to test the feasibility and effectiveness of the algorithm.