請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/50520
標題: | 建成環境對網約車使用之影響 The Influences of Built Environments on Ridesourcing Usage |
作者: | Shiyue Zhang 章詩悅 |
指導教授: | 林楨家(Jen-Jia Lin) |
關鍵字: | 網約車,旅運需求,建成環境,旅次發生,空間廻歸, ridesourcing,travel demand,built environment,trip generation,spatial regression, |
出版年 : | 2020 |
學位: | 碩士 |
摘要: | 網約車是近年来新興的運具,網約車公司利用智慧手機應用程式連接旅運者和私人駕駛員。旅運者通過應用程式發送訂單請求,當司機端接受訂單後,旅運者可以看到事實的車輛座標位置及車輛到達的預估時間。網約車在城市地區快速發展,但目前學界仍未釐清影響網約車使用的相關關係。本研究希望通過釐清建成環境因素對網約車旅次發生、旅行時間的影響,在學術上補足網約車使用在建成環境因素方向上的研究空缺。 本研究使用中國成都市『滴滴出行』網約車公司在2016年11月的訂單資料,以鄉級行政區爲空間分析單元,使用訂單起迄點發生率和起迄點旅行時間作爲衡量網約車使用之特性,將建成環境以5『D』變量密度、多樣性、設計、地點可及性以及到達大衆運輸系統的距離分類整理爲解釋變數,以旅行距離作爲旅行時間的控制變量。經過空間相關性檢驗,本研究資料存在空間相關性,使用空間落遲模型和空間誤差模型作爲本研究的廻歸模型。 結果驗證建成環境之密度、設計、地點可及性分別在旅次發生和旅行時間兩個方面影響網約車使用,對起點、迄點旅次發生均產生影響。建成環境影響網約車使用的特徵在平日和假日之間、平日一日之不同時段、假日一日之不同時段均存在差異性。 學術上,本研究對建成環境因素進行了系統性的整理,探討其與網約車使用間的關係,證實建成環境之特徵對網約車的使用存在顯著影響關係;本研究結果比對相關文獻,發現網約車使用和計程車使用、大衆運輸系統使用存在差異。實務上,本研究建議城市規劃者規劃城市建成環境時應注意從密度、設計、地點可及性三個方面減小網約車的出現對城市交通系統造成的負擔;建議網約車業者規劃營運策略時,以建成環境因素作爲車輛調度和動態調價的依據。 In recent years, an app-based service, which connects travelers and drivers and is called ridesourcing, provides a new choice of modern mobility. This service is provided by transportation network companies (TNCs) such as Uber from the U.S. and DiDi Chuxing from China. Travelers can use mobile apps to send travel requests to TNCs, once a request is accepted by a driver, the traveler can see a real-time updated location of the driver and the estimated arrival time. Ridesourcing services have a rapid growth in urban areas, but the determinants of ridesourcing usage is still unclear in the literature. This research aims at clarifying the impacts of built environments on trip generation and travel time of ridesourcing uses. This research used the complete services records of DiDi Chuxing TNC in Chengdu, China in November, 2016 as the study data and applied spatial regression methods to analyze the data. Using township level division as a spatial observation unit, the dependent variables include number of requests at origins and destinations (O-D) and O-D travel time. To explain the dependent variables, the 5Ds variables including density, diversity, design, destination accessibility and distance to transit are applied to measure built environment attributes, and the travel distance was used as control variable. With the help of Moran's I analyses, the observations were confirmed to be spatially auto-correlated and hence the Spatial Lag Model and Spatial Error Model were applied to regression analyses. The empirical findings of this research confirm that the built environments of density, design and destination accessibility were associated with ridesourcing uses and O-D travel time and the associations were different between weekdays and weekends and among time periods in a day. This research verified the influences of built environments on ridesourcing uses in an overall perspective and provided an original evidence of the influences. To compare with other transports, ridesourcing uses is different from taxi uses and public transportation uses. In practice, this research suggests urban planners about how they can decrease ridesourcing-caused traffic congestions by changing the attributes of density, design, destination accessibility, and gives TNCs a reference when formulating vehicle scheduler and dynamic pricing strategy. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/50520 |
DOI: | 10.6342/NTU202002917 |
全文授權: | 有償授權 |
顯示於系所單位: | 地理環境資源學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
U0001-1108202012020400.pdf 目前未授權公開取用 | 12.14 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。