Summary: | 碩士 === 國立臺灣科技大學 === 電機工程系 === 95 === Vehicle routing problems (VRP) plays an important role in business. Its goal is mainly to search for the minimum cost and to satisfy the demands of the customers when the demands are known. VRP which contains backpack problem and travel salesman problem is a NP-hard problem, and it has the restriction of vehicle’s capacity and the largest distance. In our research, we adopt the type of single-depot, and the vehicles set out from the depot and then get back. Ant colony optimization (ACO) is a new heuristic algorithm which has been successfully applied to solve combinatorial optimization problems. ACO mainly depends on a dynamic (pheromone) parameter and a static (distance) parameter to search for optimal. However, it still has some deficits such as stagnation behavior and premature convergence. The deficits will be more evident when the complexity of VRP increases. In the thesis, vrpnc1 ~ vrpnc10 are used as our research instances. We proposed the methods of lower pheromone bound and speed parameter and random mechanism to effectively overcome these deficits and then cooperated with classical heuristic to improve the search capability of ACO.
|