| Summary: | 碩士 === 國立臺灣大學 === 工業工程學研究所 === 101 === Selective Mulit Depot Vehicle Routing Problem with Pricing, SMDVRPP, is a derived problem of Classic Vehicle Routing Problem. This problem origin from the process of recycle enterprise setting fleet to buy recyclable product owned by the dealer. It is defined as a benchmark problem by theacademic and use integer programming and Tabu Search, TS, to solve this problem, because the quality of solution is still not enough. Our research propose two kind of meta heurisitc algorithm which are genetic algorithm, GA, and ant colony optimization, ACO, algorithm try to find the so far best solution better than ever. The concept of GA is through the evolution of chromosome. It will eliminate the chromosomes which have lower fitness value to get the best chromosome. The concept of ACO is by imitating ant foraging behavior and leaving pheromone at the road, through the pheromone value to find the best solution. Besides, our research are also propose chromosome encoding and decoding way which is suitable of SMDVRPP problem. For verify quality, we connectedacademic to obtain 40 different size benchmark problems for comparison. In this study, the genetic algorithm initialize solution for this problem to compare the results of the two heuristic algorithms, the evidence shows it is effective. And compare the result among the integer programming, Tabu Search, GA , ACO, finally we found the ACO algorithm can obtain a best solution within a certain period of time.GA although the quality is slightly lower thanTabu Search, but it can get solution faster.
|
|---|
