Application of genetic algorithms for Mass Rapid Transit System Train Scheduling Problem

• Main Authors: Hung, Chen-Yu, 洪晨祐
• Other Authors: Wang, Jin-Yuan
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/38939c

碩士 === 國立交通大學 === 運輸與物流管理學系 === 102 === Train scheduling is one of the main tasks for mass rapid transit operations. Taipei Rapid Train Corporation, for example, needs to spend days to make up a new schedule whenever the operation parameters are changed. Thus, an efficient and effective train scheduling system is in need. This research proposed a genetic algorithm based train-scheduling algorithm for assigning a sequent ordered tasks to each train. This algorithm accommodated operations characteristics and various practical constraints. A real world data from Taipei Metro Wenhu line was used for testing purpose. The testing results indicated that the proposed algorithm is capable of producing feasible train schedules. Furthermore, the sensitivity analysis indicated that the impacts of number of iterations and mutation rate are not significant