Research On Optimization Of Terminal Logistics Path Based On Improved Ant Colony Hybrid Algorithm

With the rapid development of e-commerce and new media platforms,online shopping has been integrated into thousands of households and become an inseparable part of people’s production and life.The related problems of logistics and distribution system have been paid more and more attention.Having a good logistics distribution system is not only the pursuit of enterprises but also the pursuit of customers.By summarizing the literature at home and abroad,it is found that the existing optimization of logistics distribution path mostly focuses on vehicle routing problem(VRP),and for the customer-oriented interaction link of "last kilometer",it is more to optimize and imagine it at the theoretical level by using SWOT analysis.Based on this,this thesis abstracts the "last mile" distribution problem into multiple traveling salesman problem(MTSP),The ant colony algorithm and the particle swarm optimization algorithm are the two kinds of intelligent algorithms to solve the shortest route of express delivery.Because the multi-traveling salesman model has complex constraints,and there is no public data set to test the performance of the algorithm,this thesis starts with the traveling salesman problem(TSP),and uses ant colony algorithm to solve the basic TSP under the idea of K-means clustering.The introduction of K-means algorithm is inspired by the fact that multiple travelers share the same task in the multi traveling salesman problem.There are too many target points in large-scale TSP.In order to avoid excessive line crossing,the overall TSP is divided into several sub TSPS by K-means algorithm,so that the complex large-scale TSP can be divided into multiple tasks.The idea of divide and conquer is used to plan the route of each individual sub task,By testing on the public data set provided by the traveling salesman,the operation efficiency of the algorithm can be improved,which provides a strong support for the feasibility of multi task allocation in the multi traveling salesman model.Secondly,considering that when optimizing the courier’s distribution route in the real scene,the complexity of the model is greater and the constraints are more,and the ant colony algorithm can not solve it better due to its own limitations,so we choose to integrate particle swarm optimization algorithm into ant colony algorithm to improve the performance of a single algorithm and complement the advantages of the two intelligent algorithms.Through the verification of public data sets,the fusion algorithm has been greatly improved in operation efficiency and solution accuracy compared with the two single algorithms,which provides a better solution method for the final optimization of terminal logistics distribution model.Finally,the experiment of solving the end logistics distribution model by hybrid algorithm continues the idea of reducing the size of TSP by K-means clustering algorithm.The terminal logistics distribution problem is modeled as MTSP.In the multi traveling salesman problem,K-means algorithm is no longer needed to divide tasks,and m traveling merchants can share n target points to realize multi task allocation.The idea of adding timeliness constraints to the model is the same as the purpose of K-means clustering.It not only splits the task,but also balances the number of target points.The fusion algorithm is used to optimize the model.By reasonably constraining the timeliness,the task balance can be guaranteed in the case of different number of couriers.At the same time,the total journey of M couriers is reduced to the greatest extent and the distribution efficiency is improved.