The Vehicle Routing Problem Research Based On Improved Ant Colony Algorithm

VRP (Vehicle Routing Problem) is critical in the logistics scheduling system. Its role is to select a reasonable transport path for delivery vehicles. A good logistics scheduling system not only improves service quality and resource utilization, but also enhances customers'satisfaction on business services, enhancing the competitiveness of enterprises.Ant colony algorithm is an algorithm simulating evolutionary which is inspired by real ant colony foraging behavior. It has achieved effective results in solving combinatorial optimization problems such as VRP. The main purpose of this study is to improve ant colony algorithm and apply to solve the VRP. Those improvement strategies for ant colony algorithm are as followings:Firstly, initial value ofτmax is improved. In the thesis, Greedy algorithm is used to generate initial feasible route as to get the total length of the initial path. Then, pheromone update is improved. In MMAS algorithm, only optimal pheromone on the path is updated. But, in this thesis, the pheromone is updated in the optimal solution or the closest to this. In this way, sub-optimal solution pheromone on the path will be accumulated more quickly than others, effectively avoiding the premature convergence. Finally, making pheromone evaporation factor adapt by itself can avoid the decreasing of the convergence rate caused by the P too small, increasing the convergence rate and global search capability of algorithm.Simulation experiments show that we can improve ant colony algorithm proposed not only in increasing the performance of ant colony algorithm but also achieving good results in the VRP problems. Therefore, the effectiveness of the algorithm is confirmed.