Research And Application Of Vehicle Routing Optimization Problems Based On Ant Colony Algorithm

With the expansion of the scale, combinatorial optimization problems often feature a combination of the explosion; such problems can not use conventional methods to solve. The problem belongs to NP-Hard. Ant colony algorithm (ACO) is created by the natural world which is to imitate the process of ants searching for food. The study found that ant colony algorithm can be used to solve combinatorial optimization problems, such as VRP. Ant colony algorithm finds better solutions of strong, distributed computing, robust, easy-to-combination, etc. These performances make it have very broad application prospects. However, the ant colony algorithm has disadvantages like slow convergence in the expansion nodes and so on. What's more, the ant colony algorithm in the vehicle routing problem cannot make traveler return to repository center properly. In addition, how to make the ant colony model Algorithm join in the Software system with web application has become an urgent issue. These problems are also huge challenges that we are facing.In this paper, full analyses of the current ant colony algorithm are listed on the issue. Then we analyze the basic ant colony algorithm in detail, and then elaborated on the issue VRP and its mathematical model. The increasing scale nodes of the combinatorial optimization problems lead to deal with problem more difficult. Avoiding a long distance routes, combing a short distance routes, we can limit the input of algorithm to accelerate the convergence ant colony algorithm. For vehicle routing problem, the traditional ant colony algorithm formula will not lead vehicle return to the warehouse correctly. Then the scholars improve the algorithm to make the vehicle can return to central node. However, this method cannot determine the parameters and ruins the formula. For this situation, we improve the formula with a new probability. The vehicle returns to center according to the probability which is made by vehicle load and distance between vehicle and center. Through prove and testing we verify the correctness of this new probability formula. Finally, using web GIS (geographic information system), we fill algorithms into visual maps and publish out the service through the web to make algorithm easier for the customers' browser and other operations. Then the ant colony algorithm with a geographic information system as a smart logistics and distribution management system module, making it support for multiple users to make algorithm become a software product. The first chapter is the introduction. The second chapter we start with some basic theoretical concepts, including the ant colony algorithm, as well as the basic principles of vehicle routing problem definition and mathematical modeling. And then we review the already usage of the ant colony algorithm to solve vehicle routing problem. Then some of the geographic information system concept is introduced. The third chapter is the core of this paper which puts forward a series of improved methods about how to sub-division and resolve practical proposals. Chapter IV is the intelligent management of logistics and distribution system. Using jsp / servlet/ GIS to deploy services, and to improve the user's visual operations. Chapter V is summary and prospect which points out the future direction of research.