Optimal Design Of Multi-echelon Reverse Logistics Network

Since last decades, driven by environment protection, government's legislation, economic interests, corporate responsibility, reverse logistics has drawn more and more attention from government, manufacturer and customer. Successful cases of reverse logistics from the industry become a hot academic research and the academic research promotes the successful applications of reverse logistics. As an important part of reverse logistics, the network design problem of reverse logistics is a strategic issue and determines the operation performance of reverse logistics in largely. The existing study about network design problem of reverse logistics are only aimed at minimizing the logistics cost and few relate to multi-echelon network. Taking into account the diversity objectives of logistics network system and the complexity of the network structure, this paper study the design of multi-echelon nonprofit reverse logistics network in the context of solid waste network and the design of multi-echelon business reverse logistics network in the context of repair service network.First, this paper introduces the backgrounds and significance of study. Based on the review of literatures, the problem to be studied is proposed.Secondly, this paper studies the multi-echelon solid waste network design problem which has variation material flow. This kind of network is divided into two kinds, one is urban solid waste network and the other is regional solid waste network. For urban solid waste network, this paper studies optimal network design for municipal solid wastes which involves a number of issues, such as the selection of locations of treatment facilities and landfills, determination of the capacities and number of treatment facilities and landfills, waste flow allocation between these facilities. An integer linear programming model has been proposed. A two-phase tabu search algorithm has been designed to solve the model. A set of data has been randomly generated to test the algorithm. The computation results show the algorithm is very effective and efficient. For regional solid waste network, with respect to multiple type of facilities, multiple waste flows and the waste production amount which are treated as fuzzy numbers, a fuzzy chance constrained programming model is presented to arrive at the optimal configuration of the network system components. The way to convert chance constraints into theirs respective crisp equivalents is also discussed. An example is provided to show the feasibility of the model.Thirdly, this paper studies the solid waste network design based on tradeoff between cost and distance. This kind of network covers a part of city. The distance between facility and dwelling zone is used to measure the positive or negative impact imposed by facility on dwelling zone. Two multi-objective models are proposed according to the two-layer network and three-layer network. Using multiplication- division method and fuzzy method, the two multi-objective models are transformed into single objective models individually. So, a Pareto solution can be obtained. An example is provided to show the feasibility of the models.Fourthly, this paper studies the solid waste network design problem based on tradeoff between cost and negative utility. A negative utility function is defined to measure the impact imposed by facility on dwelling zone. A multi-objective model is developed. Two objectives, minimization of the overall cost and minimization of negative utility are addressed. An evolutionary algorithm is designed for solving the model. So, the Pareto set can be obtained which can be used to design different solid waste network. The above model is extended to multi-period case. The new model allowing for building new facility to expand the capacity of old network is proposed. A hybrid multi-objective evolutionary algorithm to solve this problem is developed. The computation results show the proposed programming model and algorithm are effective approach for the problem.Fifthly, this paper studies the after-sale service network design problem. The coverage is used to measure the service level. For the two-layer network, a two-objective model is proposed. A multi-objective evolutionary algorithm to solve this problem is developed. In section 5.2, to address the problem of location, inventory, transportation and service coverage in reverse logistics, a nonlinear integer programming model is proposed. A hybrid intelligence algorithm based on"greedy algorithm"is developed which can efficiently solve the problem. In section 5.3, a three-objective model is proposed to design a four-layer network. The objectives considered in the model are minimization of the cost, minimization of the responsive time, maximization of the population covered by the facility. Based on the fuzzy satisfactory level of objectives, the model is transformed into a single objective model which maximization of the total satisfactory. An example is provided to show the feasibility of the models. In section 5.4, taking the customers choice of facility into account, a tree-objective bi-level programming model is developed to design a four-layer network. The down level of model is customers choose the facility and the upper level is Enterprises minimization of cost, maximization of coverage.Finally, the paper concludes the research and presents the prospect of this topic in further study.