近年來計程車空車率逐年攀升的情形逾加嚴重,截至民國95年為止,台北地區之計程車空車率已高達百分之七十點四。不同於過去計程車相關研究所著重之議題,如費率票價之制定、管理政策、計程車共乘、市場供需平衡與降低空車率等議題,本研究希望能在現行高空車率之情形下,為了減少計程車業者無謂的繞行,從有效接載乘客的觀點進行候客站點停等的選擇,並選取較有可能成功載客之時空區位進行候客。 根據過去文獻可以得知土地使用、時段分類等計程車候客區位上之時空特性因子將會影響計程車業者於各站點之載客次數,本研究將上述時空特性因子置入多元迴歸模式之中,並再加入替代運具之相關影響因子,求出影響載客次數之多元迴歸式。自研究成果得知,各因子中以替代運具班次數對於計程車載客次數的影響最為劇烈。 最後本研究將前述之時空特性因子代入台北市各分區(東、南、西、北)之中進行候客站點之選擇模擬,並從模擬結果可以發現,業者若自東、西區出發,則其營運方式多採取於同分區內進行候客,而自南、北兩區出發之業者則大多會採取跨區候客之形式。綜合上述研究成果可以發現,時段與區位之因素確實對於計程車之營運成效有明確影響,故本研究應可作為計程車選擇招呼站點候客上之參考。
The vacancy rate of taxi continued to raise in the last few years; up to 2006 the taxi vacancy rate of Taipei Metropolitan has already exceeded 70.4%. Different from the past researches focusing on issues of taxi pricing, administrative management, taxi pooling, supply-demand and vacancy rate, this study aims to compare loading strengths of taxi stands with differential ridership possibility and to characterize factors affecting that possibility. According to the results of previous empirical studies, the time-location factors, such as landuse types and time period, usually affect the mode choice of urban users. This study further examined the interaction between temporal and locational factors, as well as the relationship between factors and passenger loads by applying regression models. The results show that the other factors, such as services of alternative modes, also play a significant part of affecting passenger loads. Thereafter, this study applied a searching algorithm of modified minimum spanning tree to simulate the best queuing locations for optimizing passengers carrying in different districts of Taipei city. The simulation results reveal that if taxis’ origins at East or West district, they tend to operate in an intra-district pattern; however, those who originate at North or South district prefer to operate crossing districts. In summary, this study shows that time-location factors indeed significantly affect the operating outcome of taxi, and the simulation results can be used as reference for taxi choosing stands by time-location considerations.