The Research Of Vehicle Scheduling Problem With Full Loads

Vehicle scheduling problem with full-loads has long been a hot research issue in logistics & transportation industry. As the transportation cost occupy a great proportion in the circulation cost of goods, a small decrease of it would cause a great deal savings. A scientific and rational vehicle scheduling method will not only reduce the cost, but also increase the circulation speed, improve the operation efficiency and raise the service level. However, there exist many constraints and optimization targets, which cause it difficult to solve. So, how to solve the VSPFL with multi-constraints is the topic this paper focuses on.This paper begins by introducing the background and the actuality of the research of VSPFL, reviews the literature in this field and sums up the shortcomings existed in the research. The content and the frame of the research are also put forward in this part.Next, the VSPFL is described in detail, including its definition, objective and sorts. On this basis, two kinds of solving methods are introduced. One is the math method, such as branching and column-generation. The other is the heuristic method, such as saving and gene algorithm. Afterwards, an improved method based on saving is developed which aims to minimize the vehicle cost. After summarizing three basic route structures and adding both fixed and variable costs of vehicles into saving calculation, a optimized route was got by choosing the connecting technique like saving maximization or opportunity-saving maximization. It proves to get an optimized result via the example.Furthermore, Constraint Programming, a new idea used to solve VSPFL, is put forward. Its character, essential, modeling procedures and skills, solving frame and key issues are all presented in this part. Dispatcher, a software tool which applied the CP thought and used to solve the large-scale VSP is also referred in this part.Then, a VSPFL with multi-constraints in practice is discussed. After the description of this problem, we analyzed the condition carefully, and construct thesolution to it. In this part, we model using Dispatcher, which can easily add kinds of constraints into it;we solve using two-phase method and CP thought. The solution proves to be successful after the analysis of its performance.Finally, the paper summarizes the main achievement of the study and point out the future study direction.