| Summary: | 碩士 === 國立海洋大學 === 航運管理學系碩士在職專班 === 90 === Abstract
The Vehicle Routing problem has been an intensive research area during the past several decades, but typically it has been assumed that there is only one type of vehicle involved. In fact, there are many complicated factors to be considered, such as whether highways or downtown streets are employed, the traffic flow, the characteristic of the goods, customer demand, and many others. How to select an appropriate type of vehicle to meet customer needs is a rather complicated problem.
The main purpose of this study is to discuss the vehicle routing problem with hard time windows (VRPHTW), and to develop an efficient algorithm to assist managers with route planning and to reduce transportation costs. The heuristic algorithm developed in this study includes three parts: first, a revised savings algorithm is used to build up an initial solution; second, a reduction modular is implemented to reduce the total number of vehicles; third, the classical route improvement heuristic is used to shorten the distance of all routes.
Our algorithm was programmed in C++ language, and run on a Pentium Ⅲ 800MHz PC. Computational testing on 56 benchmark problems by Solomon has shown that the overall solution quality of our algorithm is not competitive with all existing heuristics. In general, the percent of error possible between the optimum number of vehicles and the distance is within about 20%.
Due to the use of the traditional heuristic algorithm, the results cannot avoid being only a local optimum. It is suggested that the future research be combined with artificial intelligence, such as TS (Tabu Search) and, TA (Threshold Accepting), in order to achieve better results.
|
|---|
