Research On Theory And Method Of Vehicle Routing Problem In Supply Chain

With the rapid development of economic globalization and electronic commerce,there is a clear trend towards a more aggressive global competition.Enterprises must therefore effectively optimize the supply chain,in order to satisfy the current market needs at the lowest possible production cost.This approach towards supply chain management has become one of the core realities that enterprises must face and address.By considering that the base commodity cost cannot be further reduced,logistics scheduling becomes a key component of the supply chain system with the potential of cutting down the overall production cost without hampering the quality of the final commodity.It therefore becomes essential to research optimal scheduling routines for logistics transportation within the supply chain system.Also,there is a great practical significance in research that addresses the vehicle routing problem in supply chain(VRPSC).For scenarios consisting of a different number of suppliers,manufacturers and stockists,considering different constraint conditions,several VRPSC mathematical models are established in this paper.Furthermore,according to the characteristics of these models,several algorithms are constructed towards the realization of efficient solutions.The main contributions of this thesis are as follows:(1)A chaotic fireworks algorithm for the multi-type vehicle routing problem in supply chain is proposed.For scenarios consisting of one supplier and multiple stockists,considering multi-type vehicles,capacity of the vehicle,fuel consumption of the vehicle,maximum transportation distance of vehicle,and aiming to minimize the fuel consumption and the delivery distance,a model of multi-type vehicle routing problem in supply chain(MTVRPSC)is established and a chaotic fireworks algorithm is proposed towards addressing it.The proposed algorithm is based on the basic principles of the fireworks algorithm,and proposes encoding and decoding schemes aimed at shifting the continuous space and the discrete space of MTVRPSC.The approach also utilizes the penalty function and vectorial comparison to deal with the fitness function.The proposed algorithm also adopts the chaotic initialization strategy and chaotic search strategy to enhance the convergent effect.Experimental results show that:the proposed algorithm has the strong optimization ability and stability in addressing MTVRPSC.(2)A quantum fireworks algorithm for the vehicle routing problem in supply chain with multiple time windows is also proposed.For the state of multiple vendors and multiple retailers in supply chain,considering multiple time windows,multi-type vehicles,capacity,transportation cost,product cost and other factors,a model of vehicle routing problem in supply chain with multiple time windows(VRPSCMTW)is established,which is solved by a quantum fireworks algorithm.The proposed algorithm takes the fireworks algorithm as its core,and proposes a quantum initialization strategy and a quantum evolutionary strategy to improve the ability of global search.Additionally,the algorithm adopts the largest order value encoding and multi-type vehicle decoding schemes in switching between the continuous space and the discrete search space.Experimental results show that the proposed algorithm is sufficient in solving VRPSCMTW effectively.(3)A discrete bat algorithm for the vehicle routing problem in supply chain with transferring is proposed.Based on the supply chain with multi-supplier and multi-retailer,considering the distribution center which is responsible to transferring the products,a model of the vehicle routing problem in supply chain with transferring(VRPSCT)is constructed and a discrete bat algorithm is proposed to solve it.The proposed algorithm presents the discrete bat position,discrete bat velocity and a series of discrete bat updated operators.Further,the proposed algorithm designs the corresponding encoding and decoding schemes for VRPSCT.Finally,experiments are used to verify the performance of the proposed algorithm and its effectiveness for solving VRPSCT.(4)A two-level discrete bat algorithm with variable neighborhood search for the vehicle routing problem in supply chain with manufacturing is proposed.Considering a two-level vehicle routing problem in a three-echelon supply chain with multiple suppliers,multiple manufacturers and multiple stockists,and aiming to minimize the purchasing,production and transportation costs,a model of the vehicle routing problem in supply chain with manufacturing(VRPSCM)is proposed.Additionally,this work constructs a two-level discrete bat algorithm with variable neighborhood search for solving VRPSCM.The proposed algorithm presents a definition of two-level bat position and introduces the corresponding operations of the bat algorithm.In addition,a variable neighborhood local search is proposed to enhance the optimizing capability of TLBAVNS,which combines 0-1 search,1-1 search with 2-Opt search.The experiments have shown that the proposed algorithm can effectively solve the instances of VRPSCM within a suitable time and significantly outperforms all the other alternatives in most of the cases.(5)A two-level chaotic discrete bat algorithm with variable neighborhood search for the vehicle routing problem in supply chain with complex products is proposed.Based on a two-level vehicle routing problem in a three-echelon supply chain,considering the fact that a complex product needs multiple materials to be manufactured,and aiming to minimize the purchasing,production and transportation costs,a model of vehicle routing problem in supply chain with complex products(VRPSCCP)is proposed,and a two-level chaotic discrete bat algorithm with variable neighborhood search is constructed to solve it.The proposed algorithm takes the two-level bat algorithm with variable neighborhood search as the core,and adopts a discrete chaotic initialization strategy to enhance the convergence effect.Firstly,parameter experiments are designed to analyze the effect of kernel parameters for the proposed algorithm so as to determine better parameters.Furthermore,the experiments have shown that the proposed algorithm can effectively solve VRPSCCP and the discrete chaotic initialization strategy significantly improves the convergence of the algorithm.Lastly,according to t-test experiments,there is a significant difference between the proposed algorithm and comparative algorithms.Finally,the thesis is summarized with conclusions drawn.An outlook on the future work is further presented.