| Summary: | 碩士 === 中華大學 === 運輸科技與物流管理學系碩士班 === 94 === In the past, the researches about bus operation have had the same shortage, which was always without complete and reliable OD Table. Fortunately, after the popularity of Taipei Smartcard, the data of rider pay can be recorded by Smartcard at the same time. Gradually, the public transportation has become a large database industry, which has many useful concealed information. In addition, data mining is also a good technique to analyze the stored data in large databases to discover potential information and knowledge.
This research constructs real passenger OD Table by Taipei Smartcard System’s data and Taipei County e-bus System’s data. As well as explore real OD Table to generate bus operation strategies. This research discusses the short-turn service route and express service route. It applied Data Mining technology about clustering and association rules to figure out optimal short-turn service route and optimal express service route, with the objective to save maximum the sum of operator’s cost（including traveling time cost and distance cost）and passengers’ travel time cost（including in-vehicle time cost and waiting time cost）. A case study by 802 route shows an optimal condition.
By this research, the result shows that the best improving effect of short-turn service is the afternoon peak time period, and the best improving effect of express service is the morning peak time period. It also reflects that operators should base on passengers’ demand to adjust their operational strategies. Overall, the effect of short-turn service is better than express service. However, the performance of optimal short-turn service route solution surpasses the present short-turn service. By the sensitive analysis, excluding from the cost of passenger’s waiting time, we could save more total cost if other factors’ costs arise. It implies that providing customized service not only the passengers may get more suitable bus service but also the operator may save more cost.
|
|---|
