Study On Integrated Optimization Of Vehicle Filling Problem And Vehicle Routing Problem

With the logistics status in society becoming more and more important, optimization problems in the logistics system has become a hotspot. Both vehicle filling problem and vehicle routing problem belong to the optimization problems in logistics operation and management, besides they are two core issues. They are directly related to the entire logistics system efficiency and effectiveness, so they must be studied comprehensively from theory to application.Vehicle routing problem, VRP for short, and vehicle filling problem, VFP for short are interrelated closely. For example, the loading sequence in VFP is directly determined by VRP while the solution process of VRP must take vehicle loading weight and space of VFP into accout. That is why VFP and VRP should be considered together.By finding the entry point for the VFP&VRP combined optimization, this paper incorporates the two problems, builds the integrated and optimized model and proposes a combined algorithm which is suitable for the model. In the loading part we consider the cargo fragility, loading stability, oriented loading, balance weight distribution, last in first out and other practical constrains. This not only can assure the transportation safety and prevent low efficiency and cargo damage because of cargo overturned but also ensure that the solution for VRP is feasible, that is cargoes at the same route can be loaded completely.The combined algorithm is made of modified loading heuristics(MLH) and modified C-W based ant colony algorithm(MCW-ACO), that is searching the VRP solution by MCW-ACO while checking the feasibility of VFP by MLH. MLH adopts the strategies of searching and merging residual spaces which makes room for each vehicle fully utilized, thus improving the loading efficiency. Based on the standard ACO, MCW-ACO introduces the thought of saving algorithm and modifies the transition rule and pheromone update formula, which improves the searching speed and results. For proving the feasibility and correctness of model and algorithm, the paper gives the Bench-mark test result. It shows that the CA can solve the large-scale VFP&VRP and get satisfactory solutions.