Research on VRPTW optimizing based on k-means clustering and IGA for electronic commerce

Abstract

Vehicle route problem with time windows of logistics distribution is the important step optimizing logistics distribution and indispensability content of electronic commerce activity. For VRPTW optimization under electronic commerce is a special problem that includes many aspects, hybrid strategy is usually introduced to classify and optimize route by two artificial intelligent methods. Therefore, the improved two-phase algorithm needs to be adopted to get solutions. Namely, the customer group can be divided into several regions using k-means algorithm in first phase. And in every region it can be decomposed into small scale subsets according with some restraint conditions using scan algorithm. In second phase, it is route optimization problems of several single TSPTW model. Therefore, the study proposes the improved genetic algorithm. Improved partially matched crossover operators can avoid destroying good gene parts during the course of crossover so as that the algorithm can be convergent to the optimization as whole. According to the traditional genetic algorithm shortcomings of slowly convergent speed, weakly partial searching ability and easily premature, the study adopts the strategy of protecting gene as whole, introduce adopts 2-exchange mutation operator, combine hill-climbing algorithm and construct new genetic algorithm on basis of establishing model of optimizing vehicle route with time windows. New algorithm offers a very effective method to solve problem of distribution vehicle schedule with time windows through the test.