Research On Improved Genetic Algorithm For Solving Optimization Of Logistics Distribution Route

With the continuous improvement of people's living standards,online shopping has become a very popular consumption in our daily life.The popularity of online shopping provides a better market prospect,it raises new issues and challenges as well.Facing with a mountain of goods,every courier company wants to use a shorter time and more effective way to handle all goods.Rational planning of each vehicle travel route not only can conducive to increasing urban traffic throughput,but also can save costs for enterprises in order to create considerable economic value.Since the multi-collar car distribution optimization problem can be decomposed into a single car route optimization according to certain policies,this thesis aims at discussing and studying the problem of logistics and distribution route optimization by a single car.The single vehicle routing problem,namely the TSP problem(Traveling Salesman Problem,TSP).As combinatorial optimization and NP-complete problem,we often take a lot of time and effort to solve TSP,and most results are unsatisfied.When it comes to the shortcomings of Genetic Algorithm of slow convergence and easy to fall into local optimum,this thesis puts forward an improved algorithm GBLSA(Genetic Based on Link-State Algorithm).This thesis gives a detailed introduction of TSP and Genetic Algorithm about relevant theoretical knowledge,trends and research directions.It discusses the methods and ideas of the traditional genetic algorithm for TSP,aiming their own shortcomings put forward a new improved method.This method combines the Link-State Algorithm and improves the three operators of Genetic Algorithm.The main research is as follow: First,generating a set of population randomly and choosing the initial one from it,then giving every individual a weight according to their rank.Then,we make full use of the ability of link-state algorithm for global optimization of populations into cross-operation aiming at improving the individual generation by generation.Finally,we execute the mutate-operation base on changing the value of gene.In addition,the Simulated Annealing and improved adaptive probability is introduced,which can retain talented individualseffectively and solve the shortcomings of local optimum.The experiment suggested that this algorithm runs better than other traditional methods in searching efficiency and performance.