| 摘要: | 近年來隨著台灣地區大專院校數量的增加，相對地學生人數與之劇增，各校間教學合作日益頻繁，校際間之選授課行為也因應而生，台灣聯合大學系統即為最佳範例。現行之跨校選授課專車，目的在於服務台灣聯合大學系統之教職員與學生，但跨校選授課專車之班表為人工經驗排班方式進行規劃，因缺乏系統性分析，故無法完善地服務搭乘需求。緣此，本研究針對跨校選授課專車之特性，以學校立場為考量，構建一系統最佳化模式，以期能提供學校方面作為有效規劃之輔助工具，並協助決策者有效地進行規劃。 本研究利用網路流動之技巧與數學規劃方法，建立一考量多對多起迄需求之跨校選授課專車排程模式，以最小車輛數及最小總加權營運時間為目標，於網路設計上，加上時窗限制以滿足實務之營運條件。本研究之模式為含額外限制之多目標整數網路流動問題，屬NP-hard問題，為有效率地求解大規模之問題，本研究利用問題分解策略，且配合CPLEX數學規劃套裝軟體，發展一有效之啟發式解法。最後評估本研究模式與啟發式解法之實用績效，以台灣聯合大學系統為範例測試，並針對不同之參數進行敏感度分析與方案分析，其測試結果良好，最後提出結論與建議。 In recent years, with the increasing of universities in Taiwan, the number of students rises as a result. With this situation, the behavior of having inter-school classes appears. The University System of Taiwan is the best example. Nowadays, the inter-school bus is set up to service the teachers and students of The University System of Taiwan, but the schedule is manually performed by planning personnel with experience in current practice. Without a systematic analysis, it cannot satisfy the demands perfectly. Therefore, in this research, consider the properties of the inter-school bus problem and based on the perspective of school planner, we develop a system-optimized model. The model is expected to be an effective tool for the school, also, it can help planner to solve problems. In this research, we employed network flow techniques and mathematical programming method, and build an inter-school bus scheduling model with many-to-many OD demand. The minimum number of vehicles and the minimum total weighted operating time is our target. In the network design, adds time window limit to meet the practical operating. Model is formulated as an integer multiple commodity network flow problem, which is characterized as NP-hard. To efficiently solve the problems whose sizes are expected to be huge in real practice, we developed a heuristic algorithm by use the problem decomposition strategy with CPLEX software. In addition, to evaluate the performance of inter-school bus model and solution algorithm practice, we take the University System of Taiwan as an example. Performing sensitive and scenario analysis for different parameters, and the result of the model test shows good. Finally, conclusions and suggestions are given. Key words: Inter-school bus, Time-space network, Heuristic, Multiple commodity network flow problem. |
