Study On City Joint Distribution Problem And Algorithm Considering Simultaneous Home Delivery And Customer's Pickup Services Under Online Shopping

With the improvement of the e-commerce environment and supporting technology,online shopping which works as a new shopping way has become a favor due to its convenience and the pleasant experience when surfing the Internet.Working as the supporting stage of the e-commerce supply chain,city logistics is experiencing a sustainable boom along with the rapid development of online shopping market.However,different to traditional city distribution,city logistics under online shopping is characterized by spatial locations,personalized demands with small batchs and uncertain factors,which have made city distribution the most expensive and polluting but least efficient stage of the e-commerce supply chain.Besides,traditional extensive operation mode has caused serious waste of resources and environment problems such as traffic congestion and noise pollution etc.However,the current studies on city logistics under online shopping paid more attention to the distribution mode,management strategy and service system.In-depth consideration of the characterstics brought by online shopping were seldom tackled,which cannot benefit the practice.To provide personalized services considering the cost and environment effects,joint distribution with simultaneous home delivery(HD)and customer's pickup(CP)is the case in this study,the dissertation does the research as follows:(1)We first focus on the operation of a single logistics operator,to provide personalized distribution service economically,bi-level multi-sized terminal location and routing problem with simultaneous HD and CP is studied considering service options,terminal sizes as well as bi-level routing strategies.In this research,partitioning strategy is first employed considering the distribution characteristic of the cutomers under online shopping,and a service selection model based on the pickup facility's location is then proposed to describe the relationship of the two service options.Besides,an improved continuous approximation model is proposed to describe the vehicle routing problem within customer groups.By analyzing the characterstics of the model,a simulated annealing genetic algorithm is designed to provide the solution method.Finally,the validity of the proposed model and algorithm is verified based on extensive experiments.(2)Based on the proposed distribution model with simultaneous HD and CP services,mult-agent joint distribution problem with terminal sharing is studied considering the ecomomy and feasibility when implementing.In this problem,pickup facilities are opened to all agants,each terminal is available to serve the customers which belong to different agents and can be visited by more than one agent.For the considered problem,a location-routing model with simultaneous HD and CP services based on terminal sharing is established.A hybrid evolution search algorithm is then desighed to provide solution method.In this method,local search is adopted to improve the new generated solutions during the evolution process.To obtain high quality initial population,effective solution and population initialization heuristics are proposed.Besides,advanced individual evaluation,crossover and mutation operators are designed to enhance the evolution and search ability.Comparsion experiments based on a real-world instance indicate the validity of the proposed model,and the performance of the designed algorithm is verified by testing the banchmark instances.(3)In view of the validity of the two-echelon system in dealing with the city logistics,a two-echelon location-routing problem with joint distribution in the second echelon is studied.Considering the sharing of the satellites and corresponding vehicles,vehicle type combination and the coordination of the two echelons,two-echelon location-routing model based on resources sharing is proposed.Due to the solving complexity,a double-population parallel evolution algorithm is designed by analyzing the characteristics of the model.The algorithm conducts the evolution and search process by managing feasible and infeasible solutions independently,and variable probability mechanism is designed to guide the crossover between the two subpopulations and thus to exchange the gene information.Based on a real-world instance,the cost and carbon emission effects under different vehicle combinations and distribution modes are analyzed.The validity of the presented algorithm is verified based on generated instances.(4)To optimize the whole city logistics system,simultaneously considering two kinds of services and the sharing of satellites,the second echelon vehicles and terminals,a two-echelon vehicle routing model is established.Due to its high complexity in solving,based on the presented double-population parallel evolution algorithm,a hybrid multi-population parallel evolution and search algorithm is proposed.In this algorithm,more than one feasible subpopulation and one infeasible subpopulation conduct parallel evolution.The core idea of this algorithm is to improve the efficiency by sharing the current best solutions among feasible subpopulations and to expand the search space when conducting crossover between feasible and infeasible solutions by available probability mechanism.Besides,due to highly constrained,a three phase heuristic is proposed to generate high quality feasible initial solutions.Computation experiments based on a large family of instances are conducted,and the results obtained indicate the validity of the proposed model and method.