A Meta-Heuristic Method for Vehicle Routing Problem with Hard Time Window Constraints

• Main Authors: Yu-Jing Tsai, 蔡玉晶
• Other Authors: James Chien-Liang Chen
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/06187522153149675757

碩士 === 中原大學 === 工業工程研究所 === 92 === This research proposes a heuristic, Tabu-Threshold Genetic Algorithm (TTGA), to efficiently and effectively solve Vehicle Routing Problem with Hard Time Window Constraints (VRPHTW). TTGA integrates Tabu Search (TS), Threshold Accepting (TA) and Genetic Algorithms (GAs) that are the most popular generic heuristic in solving VRPHTW in recent years. The first objective is to determine the routes that minimize the total vehicle travel distances, and the second objective is to find the minimum required number of vehicles. Both objectives lead to quick response to satisfy customer demands and reduce the transportation cost. TTGA consists of three phases: initial solution construction, local search improvement, and generic search improvement. In the initial solution construction phase, enhanced Nearest Neighbor Method is used. In the local search improvement phase, vehicles reduction and Neighborhood Search modules are proposed. In the generic search improvement phase, a hybrid algorithm of TS, TA and GA is used to improve the current solution. TTGA results in good solution quality and efficiency. The average deviation of distance is less than 3.6% and the average deviation of the number of vehicles is about 11.5%, compared to the best known solutions of Solomon’s 56 benchmark instances.