Research On Vehicle Routing Problem With Pickup And Delivery In Crowdsourcing Mode

The logistics industry is an important part of the national economic system.In recent years,as the market scale of e-commerce,new retail and other related online industries has expanded,the overall scale of the logistics market has also grown rapidly.Although the logistics industry has a bright future,many small and medium-sized logistics companies are facing a severe test of survival due to the high cost of infrastructure construction and high vehicle use costs in the traditional logistics model.The novel crowdsourcing model solves the problem of large capital occupancy in the traditional model for logistics companies.The crowdsourcing model reduces the logistics construction and operating costs of enterprises by changing the organization of transportation capacity and letting idle social resources participate in the enterprise's own logistics activities.At present,this model has been successfully applied in industries such as supermarket distribution,passenger transportation,and catering distribution.The rise of crowdsourcing mode is relatively late,and the research results of classic vehicle routing problem(VRP)are also difficult to apply to the flexible scenarios of crowdsourcing mode.Therefore,designing efficient vehicle path planning for the crowdsourcing model is a key issue that needs to be solved urgently,so that the crowdsourcing model can be better applied.This paper studies the pickup and delivery problems of the car-free carrier platform under the crowdsourcing model.The car-free transportation platform uses multiple social idle vehicles as resources for pickup and delivery tasks.Each idle vehicle has its own starting point,ending point,maximum carrying capacity and service time window.It serves some customers who require pickup and delivery.This problem is an extension study of Vehicle Routing Problem with Pickup and Delivery(VRPPD),which can be described as Heterogeneous Vehicle Routing Problem with Pickup and Delivery(HVRPPD).Since HVRPPD is a complex variant of the classic vehicle routing problem,it is also considered to be an NP-hard problem.Aiming at this problem,this paper takes the minimum total transport distance as the objective function,after preprocessing and linearization process,a mixed integer linear programming model is finally established.In addition,this paper also proposes an adaptive variable neighborhood search(AVNS)algorithm,and designed six heuristic neighborhood construction operators for the algorithm.We design adaptive management mechanisms for infeasible solutions and inferior solutions respectively,so that the algorithm has adaptive characteristics.Finally,in the experimental stage,the vehicle routing problem with time window(VRPTW)benchmark test sets by Solomon and 30 sets of random examples at three scales were used to conduct multiple comparison experiments on the algorithm.The results verify that the proposed algorithm has good performance in terms of solution time stability and solution quality.