Research Of Hybrid Genetic Algorithm To Solve The VRPTW In Physical Distribution System

With the development of the market economy in our country, competition amongcompanies is becoming more and more fierce. At the same time, the methods used bydifferent corporations have been changing continually. In the past, they only focusedon the quality of their products, but now the competition is everywhere, from theproduct price, channel of distribution to customer services. And the presentcompetitive strategy is developing in the direction of the competition between supplychains. This new coming competition form will be the focus of any company withmodern logistic theory being used in the market competition. However, theshortcomings of physical distribution in modern logistic systems are exposedincreasingly. How to solve the Vehicle Routing Problem (VRP) and how to optimizethe cost in the transportation of products, so as to satisfy the customers' variousdemands and make the company profitable, this question has appealed to manymanagers and researchers. The Genetic Algorithm (GA) is an adaptive heuristic search method based onpopulation genetics. The evolutionary progress is directed by "Survival of the fittest"principle. Because GA easily drops into a "local optimum" and the quality of its initialpopulation have important influence on GA's efficiency, in general, we can't get asatisfactory result when applying the Simple Genetic Algorithm (SGA) to VRPTW. Inthis thesis, we proposed a new Hybrid Genetic Algorithm (HGA) based on Sweepalgorithm, Cheapest Insertion algorithm and IOPT mutation operators, in order to solvethe VRPHTW efficiently and effectively. In the thesis, we apply the new method to R102 dataset, and achieve a betterresult. This test is a good example to show the feasibility of our new HGA. All in all, this new algorithm is a good algorithm to solve the VRPHTW. It canget satisfying result within acceptable time. We can conclude that our method is anefficient method to solve a large scale combinatorial optimization problem.