Research on Vehicle Scheduling Problem Based on Improved Genetic Algorithm for Electronic Commerce

Electronic commerce, as a new commercial mode, has its own particularity comparing with traditional commercial activities. Logistics Company of Electronic Commerce will face with modern market of multiple batches, small volume, high time requirement and individuation demand. There are big differences from traditional vehicle scheduling in selecting distribution vehicle types, considering transportation time, routine personnel and vehicle utilizing expense, overtime expense, selecting distribution cost and route. And it is difficult in accurately measuring for complex relationship. Therefore, optimization model on vehicle scheduling based on traditional shortest vehicle route is difficult in satisfying factual requirement of logistics distribution under electronic commerce so as to high distribution cost or lose market competition ability for difficult in delivering goods on time for excessively emphasizing the shortest route. Considering the specialties of logistics distribution under electronic commerce, the traditional vehicle scheduling model is modified in order to reduce the distribution cost, objective function is modified based on minimum expense, namely, considering routine and overtime expense, vehicle waiting expense, personnel expenses and so on. At the same time, in order to improve the distribution service quality and market competition, time windows is set to hard time window. Add goods volume restriction, maximum work time, many vehicle types, vehicle load capacity restriction, maximum running distance and others in restraint conditions in order to improve the applicability and universal characteristics of model. For vehicle scheduling problem is NP puzzle, get the optimization solution through adopting improved genetic algorithm, that is, control selection strategy through individual amount so as to guarantee group diversity, improve searching ability to group and convergent speed by partially matched crossover operator and partially route reversal m- utation operator. In the final, it is proved that improved algorithm has good performance through experiment and calculation combining with concrete examples.