Applying Harmony Search to the Traveling Salesman Problem with Simultaneous Pickup and Delivery

• Main Authors: Mei-Chen Huang, 黃玫珍
• Other Authors: 胡黃德
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/09893635835041333778

碩士 === 元智大學 === 工業工程與管理學系 === 98 === Nowadays, as the increasing demand in the logistics industry, logistics operations play an important role in the logistics system. Distribution cost accounts for the large percentage of total cost in logistics activities. Hence, a well designed path planning can reduce the transportation cost. Under the limit of just visiting once for demand points, the pickup and delivery simultaneously can avoid unnecessary waste of time and cost. This study aims to combine the Harmony Search (HS) and the 2-Opt local search (2-Opt LS) to solve the Traveling Salesman Problem with Pickup and Delivery (TSPPD) problem with the objective of minimum transportation distance. Based on the tour construction (Random Generation and Savings Method) methods and encoding of solution (Order-based Encoding and Node-based Encoding), our HS can be divided into four categories and then tested their computational performances by the statistical software MINITAB. The results showed that the Node-based Encoding is better than the Order-based Encoding significantly. The computational performance of the proposed HS is tested on 300 benchmark test problems that were obtained from the literature. The computational results showed that the Savings Method with Node-based Encoding outperformed the other two competing heuristics, (i.e., TS2 and UB2), which were proposed in the literature. Hence, the proposed heuristics can provide a fast and stable, and better solution on the TSPPD problem.