Research And Application On Dynamic Logistics Knapsack Based On Swarm Intelligence Algorithm

With the great development of modern logistics industry,in response to the national intelligent development strategy,plenty of departments,traditional manufacturing industry,modern e-commerce enterprises,express logistics enterprises,third-party logistics enterprises etc.,increase the promotion of logistics intelligence projects.Therefore,the research on logistics system and logistics system optimization by scholars at home and abroad is expanding.Under this background,this paper studies the distribution optimization in logistics system.Starting from the traditional vehicle routing optimization problem,this paper synthesized the vehicle routing problem(VRP)with the logistics knapsack problem(KP),within the constraint of distribution time window of clients,forming the dynamic logistics knapsack multi-objective optimization problem(DLKMOP).Aiming at DLKMOP,this paper proposed an improved grey wolf optimization algorithm,and validated the algorithm in the logistics system simulation model.The following aspects of work had been accomplished:1.In the basic of VRP,the constraints of customer dynamic time window and dynamic logistics knapsack optimization were added to form a multi-objective optimal problem of dynamic logistics knapsack.The main optimal objectives were the distance of distribution vehicle,the knapsack value of distribution vehicle,and the customer satisfaction that is mainly determined by dynamic time window demand of the clients.2.Aiming at DLKMOP,the improved grey wolf optimization algorithm is proposed,called non-dominated sorting grey wolf optimization algorithm(NSGWO)and fast nondominated sorting grey wolf optimization algorithm(FNSGWO).The NSGWO introduces the Pareto Frontier Analysis(PFA),and the strategy of non-dominant sorting(NS)and Shared Niche(SN).The time complexity of the algorithm is high.On the basis of NSGWO algorithm,FNSGWO algorithm added fast non-dominated sorting(FNS)and congestion degree calculation(CDC)to sort the grey wolf population hierarchically.Then,the multi-objective optimization model was solved and the multi-objective Pareto frontier solution set was obtained,that was the solution of DLKMOP.3.In order to verify the validity of the model and the effectiveness and performance of the improved algorithm,a simulation model of logistics system is built to simulate the whole logistics distribution process.The experiment proves that the distribution scheme obtained by multi-objective optimization and single-objective optimization of distribution distance is better.Under the condition that the distribution distance does not change much,while the average customer satisfaction is increased by 12.3%,and the value of vehicle backpack is increased by 14.4%.At the same time,based on the simplified model of logistics system,the comparing experiments of NSGWO,FNSGWO,GWO and particle swarm optimization(PSO)algorithm proves that FNSGWO algorithm can provide a more efficient scheme for logistics terminal distribution.