| Vehicle routing problem belongs to the category of operations research, is the core work in transportation route planning and selecting appropriate delivery vehicle, improve the transport of goods in the economy plays a vital role. With the intensification of market competition and the diversification of customer needs, the most basic model of VRP is difficult to effectively guide the distribution enterprise's operation, compared to only contains basic VRP vehicle load constraint, vehicle routing problem with time Windows(VRPTW) can reflect the actual application better, therefore, from the perspective of satisfy the demand of customers and the enterprise itself benefits, research on VRP problem has attracted extensive attention of a large number of scholars. However, the current study, especially in the design of efficient algorithm still exists a lot of in-depth study of space. In this paper, combined with the actual situation of production and living, using the improved ant colony algorithm, effectively solving multi-vehicle routing problem with time Windows. The main conclusion is as follows:(1) Ant colony algorithm, the ants in the initial process of optimization, when choosing the next node according to the state transition probability to determine the probability value of alternative customer point, under normal circumstances, the ant selected node, it will show ant colony algorithm is prone to stagnation phenomenon in the process of search shortcoming. In this paper, the introduction of a random number when choosing the next node allow some ants have a certain probability of making mistakes, that is, the ant in choosing the next node will have some exploratory search and the ants after the completion of the first iteration of the operating results of genetic operators so that we can effectively expand the solution space of problem so to improve the performance of the algorithm.(2) When the vehicle routing problem with time windows solving, customer service requirements of the time window has a direct impact on the objective function, so the transition probabilities ants choose the next node only consider pheromone of path and concentration path length may not get the desired results. Based on time window has been added to the transfer probability of ant factors, the problem of the solution is closer to the optimal solution, by analysis of the results of the improved algorithm found that the improved ant colony algorithm has more advantages in solving the problem.At present, for multi-vehicles routing problem with time windows of the few studies, large capacity truck has the advantages of large loading capacity, driving distance, but it inevitably has a slow speed, low-defect distribution efficiency, and small-capacity trucks the opposite. Therefore, this paper considering different models of features, combined with the basis of different customer service requirements of the time windows on the basis of the design of an improved ant colony algorithm for multi vehicles routing problem with time window is solved, it can be seen from the results, for the same distribution tasks, the choice of mixing models and distribution program to some extent, will achieve better results. |
PDF Full Text Request