描述性选择试验设计方法研究——在出行选择行为分析中的应用
Study on Stated Choice Experiment Design——Application in Travel Choice Analysis
-
摘要: 针对如何实现合理有效的描述性选择试验设计,总结了不同试验类型(不标号试验和标号试验)的优缺点及不同试验设计方法(正交设计和有效设计)的适用范围,以通勤者出行路径和出发时间联合选择为例,利用R软件针对多项logit模型,进行正交主效应试验设计.基于得到的标号正交试验设计方案进行数据采集和模型估计,结果显示模型拟合度良好,估计结果统计性显著.Abstract: In this paper, the key considerations required in constructing stated choice ( SC) experiments were summarized and the widely used methods for SC experiment design were discussed. The difference between unlabeled and labeled design, as well as the advantages and disadvantages of orthogonal design and efficient design were compared. Additionally, the SC experiments for investigating morning commuters’ choice of departure time and route when faced with congestion charging were conducted using statistical software R. Then, a survey was conducted and data were collected on the basis of the generated labeled SC experiment. The combined choice model for departure time route was estimated using collected dataset. The estimation results showed that the model performed well and the estimation results were significant.