應用類神經網路於票價之訂定－以台灣高鐵為例

Apply Artificial Neural Network in the Decision of THSR Ticket Price

蔡富凱(Fu-kai Tsai)

指導教授 ： 王順生

朝陽科技大學/理工學院/工業工程與管理系/碩士(2009年)

摘要

西元2007年，台灣高鐵正式與大眾見面，代表的不僅只是速度；代表的是速度所帶來的現代生活新態度。高鐵加入台灣的正式運輸行列，宣佈了台灣運輸業的新革命即將開始，2008年石油高漲時，高鐵漸漸邁入穩定的成長期，卻在2008年尾聲時，遭遇國際金融海嘯，造成公司紛紛倒閉，高鐵主要顧客族群以高級商務人士及科技新貴為主。油價在此時也開始回穩，公路運輸業也開始大打價格戰，造成高鐵以橘藍雙色折扣應對，價格的調整是一個重要的課題。本研究主要透過倒傳遞類神經網路求解列車站間訂價折扣問題。
本研究將以起迄站差別折扣為研究方向，追求營運收益最大化時，各站間的最適折扣組合，將透過倒傳遞類神經網路求解，以下建議提供給台灣高鐵作為參考：
1. 平日期間採優惠較多的策略，但其尖峰時段例如上下班時間取消優惠。假日或國定假日（例春節期間）優惠少或是不優惠的情況。
2. 列車營運情形需隨時監控，提高座位利用率減少運輸成本，降低故障率，在不影響高鐵行駛的安全性及服務品質之下，適度刪減人力避免過度支出。

關鍵字

台灣高鐵 ； 倒傳遞類神經 ； 折扣

並列摘要

Taiwan High Speed Rail (THSR) started to operate in 2007. It represents not only the new era of modern transportation system, but also a solution of time-saving transportation substitution. However, THSR faced a difficult financial situation in 2008’s international finance tsunami. THSR had provided several discount policies in ticket price, but those policies seemed in vain. This research studied an alternative discount policy by Artificial Neural Network. Back-propagation neural network was applied to maximize operating income with the assumption that different stations should have different discounts. Factors including prices, visitors, mileages and transportation capacities were considered. The result shows that the ticket prices between weekdays and weekend, as well as short and long distance travelers should be different. To encourage people to take THSR and to maximize THSR’s benefit, the prices in weekends and national holidays should be higher than those in weekdays. And, the discount for short-distance travelers should be more than that for long-distance ones. Meanwhile, different discounts among different stations have proved to be suitable. THSR was suggested to monitor the operation condition at all time to change the transportation capacity and discount rates in time.

並列關鍵字

back-propagation neural ； discounts ； Taiwan High Speed Rail

參考文獻

[20] 劉鼎康，「使用類神經網路進行垃圾郵件過濾之研究」，中原大學資訊管理研究所碩士論文，2005。

[25] 羅慕君，「短期訂單預測模型之研究—PDA產業為例」，中原大學資訊管理研究所碩士論文，2004。

[3] Shyur, H. J., Luyhoj, J. T. and Williams, T. P., “Using Neural Networks to Predict Component Inspection Requirements for Aging Aircraft,“ Elsevier Science Decision Support Systems., Vol. 30, No. 2, 1996, pp. 257-267.

[4] Kosaka, M., Mizuno, H., Sasaki, T., Someya, R. and Hamada, N., “Applications of Fuzzy Logic/Neural Network to Securities Trading Decision Support System, ” IEEE International Conference on Systems, Man, and Cybernetics, 1991. ''Decision Aiding for Complex Systems, Conference Proceedings., 1991, Charlottesville, VA, USA, Vol. 3, Oct 13-16 1991, pp. 1913-1918.

[6] Hung, S. Y., Liang, T. P. and Liu, V. W.C., “Integrating Arbitrage Pricing Theory and Artificial Neural Networks to Support Portfolio Management, ” Elsevier Science Decision Support Systems., Vol. 18, 1996.