Vehicle Routing Optimization Methods With Product Classification And Pickup And Delivery Constraints

Logistics industry as a new pillar of modern national economy,its transportation scheduling problem has been widely concerned.As the core problem of vehicle scheduling and allocation in logistics transportation,VRP has become a research hotspot in recent years.As a special case of classical VRP,VRP with constraints including Product Classification,Pickup-Delivery Simultaneously,and Time Window(PC-VRPSPDTW)effectively combines multiple practical constraints,making it more challenging and practical.In the PC-VRPSPDTW,vehicle routing planning not only restricts vehicle transportation distance,but also affects vehicle exhaust emissions and customer waiting time,thus affecting customer satisfaction with service.How to flexibly plan the transportation route of vehicles and design efficient and intelligent optimization algorithm has become the key to meet the needs of customers for different kinds of products under multiple constraints.Therefore,this paper studies the vehicle routing problem with product classification and pickup-delivery constraints based on different optimization objectives.The main research content and innovation points of this paper are as follows:1.A hybrid complementary meta-heuristic algorithm(Tabu Search-Artificial Immune Algorithm,TS-AIA)is proposed with vehicle cost,vehicle travel distance,overload part and time penalty as optimization objectives without considering green transportation.In this method,the decision variables,constraints and objectives are described by mathematical expressions under the condition of known customer demand,so as to realize problem modeling.Firstly,in order to solve PC-VRPSPDTW,an initialization algorithm based on the Earliest Time and Residual Capacity(ETRC)is designed by considering the customer’s time requirement and the vehicle’s capacity limitation,and the high quality initial solution is obtained.Secondly,the proposed hybrid complementary meta-heuristic algorithm can integrate the advantages of global search algorithm and local search strategy to achieve breadth mining and depth mining of solutions.The implementation process of the algorithm can be divided into two stages: In the first stage,the tabu search algorithm is used to optimize the optimal solution in the population,assisted by variable neighborhood search strategy to reduce the number of vehicles and reduce vehicle cost.In the second stage,the artificial immune algorithm is used to optimize the solution population,and a crossover strategy based on the customer relationship table is designed.By summarizing the degree of affinity between customers,the customers with strong correlation are transferred to the offspring,so as to ensure the transmission of excellent genes.In addition,the application of large neighborhood search can effectively improve the search ability.Experimental results show that the proposed algorithm can obtain more optimal solutions for the optimization of initial solutions and fitness,which proves the effectiveness of the algorithm.2.In the case of green transportation,a Bi-Population Memetic Algorithm(BPMA)is proposed with vehicle cost,vehicle travel distance,fuel consumption and time penalty as optimization objectives.The algorithm divides the population into high quality sub-population and low quality sub-population based on the standard of fitness.According to the characteristics of different sub-populations,different crossover and mutation rates are adopted,and different optimization strategies are selected.Bi-population collaboration selection strategy integrates elite selection,perturbation selection and tournament selection,which ensures species diversity while retaining excellent individuals.A crossover strategy based on the relationship-location table is designed,which takes into account the correlation between customers and the uniqueness of each customer,and continues the gene pairs with strong correlation.According to different subpopulations,the multi-point dual-population dynamic variation strategy adopts different strategies and realize the increase of mutation rate in the late period of population evolution.A self-directed local search strategy determines the search strategy selection based on the performance scores of the three search strategies in the early stage,realizes the integrated application of various search strategies.Experimental results show that the proposed algorithm has good effects on the optimization of fitness and subobjectives,and a better solution within the effective range is obtained.