利用智慧卡票證資料探索臺北捷運乘客回返行為之研究

Abstract

NONE

Metro networks possess many stations which travelers can freely connect between while within them. Due to the exponential complexity of accounting for every possible path a passenger may take between any two stations, the concept of feasible paths helped narrow down path selection to routes which strictly advance the traveler toward their destination and away from their origin. Travelers are often assumed to follow these feasible paths as rational actors to minimize their travel time and fare costs, along with any other utility one route may possess over another. The utility of passenger comfort, considered during crowded metro situations, is still primarily perceived as a choice between two or more feasible paths. However, travelers in the Taipei Metro will go further and backtrack to their respective terminal station before heading toward their destination to achieve this comfort. While backtracking has been considered in networks where there might be a time advantage, such as backtracking to an express service station or backtracking to a transfer station with reduced transfer times, backtracking has yet to be considered strictly for the utility of traveler comfort at the expense of longer travel times and potentially higher fares. A backtracking algorithm is developed to capture this backtracking behavior from smart card transaction data and used to train a binary logit model in an attempt to find which variables most directly influence the traveler’s decision to backtrack.