標題: 利用螞蟻群系統求解具有保養需求之公共運輸排程之研究
Ant Colony System for Public Transit Scheduling with Maintenance Requirement
作者: 陳思淵
Chen, Szu-Yuan
王晉元
Wang, Jin-Yuan
運輸科技與管理學系
關鍵字: 車輛保養;螞蟻群演算法;vehicle maintenance;Ant Colony algorithm
公開日期: 2012
摘要: 摘要 車輛常因為行駛的時間太久導致內部零件損害,而間接造成事故發生,但這些事故其實透過定期的保養就可以獲得改善,但要如何兼顧車輛的排程和保養動作。因此本研究提出利用螞蟻群演算法並加上本研究設計之方法-擾動機制,期望在求解這類問題時,可以獲得不錯的績效。 本研究利用桃園客運所提供之實際營運資料,將其分成大小範例共10個範例來測試,進行參數校估以及設計之方法的測試。經由測試的結果發現,在有沒有使用設計之方法方面,有使用設計之方法的目標值明顯比沒有使用設計之方法的目標值要來的好,且隨著測試範例越大,有沒有使用的差異程度會越明顯,間接證明出來當範例越接近實際範例,獲得改善的程度會越大,因此本研究設計之演算法和擾動機制能夠有效的解決具有保養需求之公共運輸排程問題。
ABSTRACT Because of travel for too long lead to damage the vehicles internal parts, and indirectly caused the accident, but the accident actually through regular maintenance will be improved, but how to balance the process and action platoon of vehicles. Therefore, this study proposes the use of ant colony algorithm and method of the design of disturbance mechanism, expectation in solving this kind of problems, can obtain good performance. This study using the real operating data provided by Tao-Yuan Bus Transportation Company, divided into the size of a total of 10 example to test the model, parameter calibration and the design method of test. It has been shown, there is no use of design method, a design method of target values than did not use the design method of target values to the good, and with greater test example, there is no difference between the more obvious, indirect evidence out when the sample is close to the practical examples, get the degree of improvement will be bigger, therefore this research design algorithm and the disturbance mechanism to public transport has maintenance requirements effectively solve the problem of scheduling.
URI: http://140.113.39.130/cdrfb3/record/nctu/#GT079932526
http://hdl.handle.net/11536/72615
Appears in Collections:Thesis