Research On The Stochastic Inventory-Tansportation Integrated Optimization Problem

The coordination of all sectors and the improvement of the whole performance are the main characters of modern logistics, but the existing domestic researches are mainly about the macro-logistics, even if about the micro-logistics, they focus on the single sector only. For these reasons, this dissertation integrates two principal problems named the Inventory Control problem and the Transportation Organization problem into one problem, and then seeks the optimal solution of this combined problem instead of each single problem. Although the theoretical investigations of these two problems are relatively mature, the mixed problem, for the sake of known violation of benefit between them is more difficult and need study deeply. From these points of view, the paper addresses mainly on the feature and soling technique of the Inventory-Transportation Integrated Optimization(ITIO) problem from different decision levels in the distributing logistics system.First of all, based on the analysis and summing-up of a great number of relative existing researches, the paper presents the definition, characters and classification of ITIO problem from the macro- viewpoint, at one time points out some lacks in those studies and some potential research fields. These are helpful for the further research.Specifying the topic on the highway-distributing network with one supplier and more retailers, this dissertation studies the Stochastic ITIO(SITIO) problem from strategic, tactical and operational decision-making level respectively.When discussing the strategic SITIO problem, the paper addresses the procedure of establishing the mathematic model and designing of heuristic algorithm with decision variables as the number of vehicles and the storage capacity, and the objective is to minimize the sum of transportation and inventory average costs when the system be run in the future. Of which some skills such as the treatment of the random demand and the method of partitioning customers are new attempts for solving this kind of problem.The goals of tactical SITIO problem are to determine the distributing period, the allocation of inventory and the optimal vehicle routing. Based on a method of portioning customers, the paper first formulates the objective function, and provides and proves the low-bound of objective value; and then designs a heuristic algorithm using some known ideas of capacitated concentrator location problem, and verifies the designed algorithm to be asymptotic optimal simultaneously. Computational results show that the procedure is reasonable and effective.In view of the operational SITIO problem is faced by the system decision-maker daily, the dissertation uses two chapters to address this class problem applied in two situations respectively. One of two situations is assumed to be R-System. Through analyzing the feature of the combined problem, the paper formulates the problem with the objective of minimizing the one-period total cost, and presents the decomposition methodology. This algorithm includes two steps. The first step is to decompose the mixed problem into the stochastic inventory problem and the vehicle routing problem and to solve them separately. The second step is to obtain the optimal solution of the integrated problem by means of the dynamic adjusting algorithm. The approximate algorithm designed for the stochastic inventory problem is practical, even in other circumstance.When assuming that the practical background is D-System, the paper apply the four-stage solving process, namely 1) random customer selection; 2) customers assignment; 3) stochastic vehicle routing optimization; 4) the procedure of improving the quality of solution. Numerical results verify our soling thoughts.In summary, the dissertation not only stresses the methods for analyzing anddescribing the characters of ITIO problem, but also discusses the skill for solving them.