Research On Improvement Of Ant Colony Algorithm And Its Application

Ant colony algorithm is a heuristic optimization algorithm proposed by simulating the behavior of ants working together to find food.It has the advantages of simple optimization process and easy combination with other algorithms.Ant colony algorithm first emerged because of its successful application in the traveling salesman problem.Its appearance provides a new idea for solving combinatorial optimization problems,and it has been widely used in image processing,network routing,logistics and other fields.Therefore,the study on ant colony algorithm has very important practical significance.This paper mainly focuses on the principle of ant colony algorithm,and studies how to improve the ant colony algorithm.The ant colony algorithm has been favored by scholars since it was proposed.However,the existing ant colony algorithms still have some disadvantages,such as easy to fall into local optimal solution and slow convergence speed.In view of these shortcomings,this paper puts forward some improved ideas,and conducts in-depth research on the application of the improved ant colony algorithm to the traveling salesman problem and the dynamic logistics distribution path planning problem.The main research results of this article are as follows:(1)Aiming at the shortcomings that ant colony algorithm is easy to fall into local optimal solution,an ant colony algorithm based on multiple state transition operators is proposed.The state transition operators from the discrete state transfer algorithm are introduced into the ant colony algorithm to effectively combine with the ant colony algorithm.A variety of state transition operators can generate new feasible solutions through swap,shift and symmetric transformation according to the feasible solutions obtained in the process of ant search,which not only improves the diversity of understanding,but also effectively prevents the algorithm from stagnating in the local optimal solution.The ant colony algorithm based on multiple state transition operators has achieved satisfactory results in the application of the traveling salesman problem.Compared with the previous ant colony algorithm,the quality of solution obtained by the ant colony algorithm based on multiple state transition operators has been significantly improved.(2)Aiming at the slow convergence speed of ant colony algorithm,an ant colony algorithm based on data classification is proposed.The data classification method takes the direction of attraction as an important consideration.The data classification method divides the data to be accessed in the selected data set into two categories: candidate category and elimination category.The candidate class data is the candidate node for the next visit of ants,and eliminated data is eliminated,thus reducing search time.At the same time,the degree of attraction direction in the data classification method can effectively guide the algorithm to search for the optimal solution when the pheromone enlightenment effect is not significant.In addition,the 3-opt local search algorithm is added to further optimize the solution.The experimental results of different scale traveling salesman problems show that the ant colony algorithm based on data classification is compared with the algorithm before the improvement,and the optimization ability and convergence speed of algorithm are greatly improved.(3)The ant colony algorithm based on data classification is applied to the related problems of logistics and distribution.Considering the current development status of logistics industry,a solution to the dynamic vehicle routing problem based on the improved ant colony algorithm is proposed.In this solution,the overall algorithm flow for dynamic vehicle routing problem is divided into two stages: initial stage and dynamic optimization stage.Firstly,the ant colony algorithm based on data classification is used to construct the initial solution in the initial stage,then the dynamic information is collected in real time in the dynamic optimization stage,and the new customer requirements are inserted into the current planned path by the dynamic insertion method.Finally,the ant colony algorithm based on data classification is used to continuously optimize the current path.The experimental results on Solomon's example show that the proposed dynamic path planning scheme based on data classification ant colony algorithm can respond to the dynamic change of logistics distribution demand and provide a reasonable and efficient distribution scheme.The experimental results on the calculation example proposed by Solomon show that the dynamic path planning scheme based on the data classification ant colony algorithm proposed in this paper can respond to the dynamic changes of logistics distribution demand and give a reasonable and efficient distribution scheme.