Modeling And Scheduling Methods Of Several Location-Routing Problems In Enterprise Logistics Distribution System

Recently, Location-Routing Problem(LRP) in enterprise logistics distribution systems is the international and national research hot. Like modern logistics system with integrated objective of cost and service, the objective of LRP is to ingrate the location allocation problem(LAP) and Vehicle Routing Problems(VRP) and minimum the total cost of the system. Simultaneously, during the real operation of enterprise logistics systems, because of more and more serious competition and the uncertainty of the customer needs, the decision-makers have paid more attention to the operation efficiency and robustness of location facilities and routing distribution to adapt the dynamic environments. Therefore, the topics of this dissertation have the great theory significance and application value.Several LRPs of enterprise logistics distribution systems have been emphasized particularly on in this dissertation, especially to the middle or large enterprise logistics distribution systems. From the global view of the enterprise logistics distribution systems, LAP and VRP, the relationship of them have been explored from every level of enterprise operations, including all procedures of modeling, solving and simulating, proving etc. The detail work consists of:Firstly, the global modeling method of enterprise logistics distribution systems has been realized from the strategy view. The parameter variables and decision variables have been presented by using smoothing function of continuous approximately values theory, all location facilities and distribution routing from the manufacturing sites to the customers'sites have been considered and according to their properties, such as all kinds of the nodes of location facilities and the modes of transportations, different parameters have been selected and the general model has been modeled. Different with the other developed models by the detail operation data or based on the grid or discrete date, the innovation modeling in this distribution has been proposed by continuous approximately values thought and has some objectivity and real guideline significance.Secondly, the further research on the regional logistics distribution systems in enterprise has been developed. The prediction of the regional customers'need, the regional facilities location and the responding routing selections have been researched based on the detail operation data. With the help of many methods, including the forecasting based on the radial basis function network, the random restrictions representation of the qualitative and quantificational combination and heuristic methods etc, the facilities location have been modeled by considering the logistics service levels, the operation risk and benefit under the scope of the prediction of the customers'need. The sequence decision has been applied in the modeling procedure and the uncertainty and random of the customers'need have been considered. Concerning with the exact solving and heuristics solving methods, the modeling have the good robustness.Thirdly, LRPs in one node of enterprise logistics distribution systems have been explored, that is the coordination problem how the goods in distribution center have been located in the warehouse and handled from and into the warehouse. To the distribution or warehouse with the rapid turnover, kinds of goods and the limited resource, this is an important problem and it belongs to the operational level. Based on the operational properties of the warehouse or the distribution, the methods of the sorting storage and from-into handling goods have been summarized. A dynamic algorithm for the warehouse scheduling in distribution centers is proposed by simulating the dynamic scheduling procedure in enterprise. Time, Space and the kinds of limitations elements have been concerned and the proposed methods can express the dynamics well and realize the global optimization ability in the location problem of the warehouse. At the same time, the continuous operations have been processed perfectly and the coding and evolution of genetic algorithm(GA) have been designed to adapt the changeable spaces of the warehouse. The fields using GA to solve the dynamic optimization and the research on the combination of the exact solving and the heuristic solving have been expanded.Fourthly, according to the analysis of the references and the real survey, most of the evaluation elements on the facilities location in enterprise logistics distribution systems are qualitative. So, the evaluation decision methods of the enterprise logistics distribution systems based on fuzzy theory have been further studied. The evaluated objects can be some integrated schemes, such as the global system level, the regional system level or the operational system level. The trapezoid fuzzy numbers have been introduced to represent the indexed and the weights to unify the exact data and uncertain data of the evaluation indexes. Based on the concept of possibility degree of defining the trapezoid fuzzy numbers, the fuzzy complementary judgment matrix can be calculated and the rankings of the evaluated schemes under each criterion and the synthetically criteria are both obtained according to the decision-maker risk preference. The results can be used to help the decision procedure by decision-makers.In this dissertation, the simulations have been performed using all models and methods developed. The simulation data comes from the real operation procedures and the comparison and analysis have been obtained to demonstrate the effectiveness to support all proposed models and developed methods.