客運車輛擾動下調度系統之研究

Abstract

在公車營運上，業者對於排班與調度最大的問題，是如何利用有限的資源，即人員與車輛，對其做合理的調度，以提高服務品質、確保營運績效。在實務營運上，有很多因素會使得原定班表受到擾動，例如人員請假、車輛維修、旅客需求變動等，業者必須將班表做適當的調整。 因應班表受擾動的調整，不只影響部分的班表，還可能會牽動其他班次及其他車輛的班表，進而造成整個班表的擾動，不當的調整將會影響業者營運績效及服務水準，故當有事件發生時，如何有效地調整系統的班表，使班表擾動狀態所產生的變動最小、降低營運損失，實乃當前業者所普遍關切之課題。 本研究的目的在改進傳統實務的作法，考慮營運上資源之限制，希望以專家系統的方法為基礎，建立發展一因應班表受到擾動之電腦化排班調度系統。系統架構主要包括知識獲取、推理機、以及使用者界面三部分。首先透過訪談獲取排班調度所需的知識，再將這些知識規則與推理機制整合，最後使用者透過使用者界面可得到調度方案。 為測試系統的實用性，本研究以實際國內客運業排班之營運資料為例，作實例測試與修正。實務上客運業者可以針對其調度方案與考量之不同，設計較合適的調度策略，以求系統能符合業者的需求。

In order to improve the quality of service and insure the performance of operations, the most important thing in bus scheduling and dispatching is how to maintain pre-announced schedule. In reality, the bus company must readjust the schedule under abnormal circumstances such as the call-in-sick of bus drivers, breakdown of vehicles. Readjusting the schedule may inference the whole schedule. Inappropriate adjustment will affect the performance of operation and the quality of service, so how to adjust the schedule effectively becomes an important task. The purpose of this study is to adapt an Expert System like approach to improve traditional methods. First, we obtain the knowledge of dispatching rules via an interview with bus companies. Second, we combine these rules with inference engine. Finally, users can obtain dispatching strategies through the user interface of the system. In order to evaluate the practicability of the system, a case study concerning the operation of bus companies in Taiwan is performed. In reality, bus companies can design appropriate dispatching strategies in accordance with their needs to fill the bill.