A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Modeling Actual Dwell Time for Rail Transit Using Data Analytics and Support Vector Regression
Actual dwell time (DT) at rail transit stations is one of the most significant and useful parameters in evaluating the scheduled timetable and making any dynamic operation adjustments as necessary. It can be influenced by many factors with intricate nonlinearities that make it difficult to build an overall model. The purpose of this paper is to develop a data analytics approach to modeling the train actual DT by combining a density-based spatial clustering of applications with noise (DBSCAN) with the support vector regression (SVR) algorithm. There are three steps in this modeling approach: (1) identifying factors that influence train actual DT; (2) integrating data collected from automatic fare collection and automatic train supervision, and then classifying actual DT data by using the DBSCAN approach; and (3) establishing a nonlinear model to estimate the actual DT based on the SVR method, which is calibrated by using grid search techniques. Using a station along Line 12 in the Shanghai metro network as an example, this innovative hybrid approach delivers high quality results in that the mean squared error of the training set is 2.87 s and the coefficient of determination is 0.66. Application results indicate that the influential factors identified and the nonlinear model developed can be used to explain and predict the train actual DT well, and that the developed model can be applied to assist decision making at both the tactical and operational levels.
Modeling Actual Dwell Time for Rail Transit Using Data Analytics and Support Vector Regression
Actual dwell time (DT) at rail transit stations is one of the most significant and useful parameters in evaluating the scheduled timetable and making any dynamic operation adjustments as necessary. It can be influenced by many factors with intricate nonlinearities that make it difficult to build an overall model. The purpose of this paper is to develop a data analytics approach to modeling the train actual DT by combining a density-based spatial clustering of applications with noise (DBSCAN) with the support vector regression (SVR) algorithm. There are three steps in this modeling approach: (1) identifying factors that influence train actual DT; (2) integrating data collected from automatic fare collection and automatic train supervision, and then classifying actual DT data by using the DBSCAN approach; and (3) establishing a nonlinear model to estimate the actual DT based on the SVR method, which is calibrated by using grid search techniques. Using a station along Line 12 in the Shanghai metro network as an example, this innovative hybrid approach delivers high quality results in that the mean squared error of the training set is 2.87 s and the coefficient of determination is 0.66. Application results indicate that the influential factors identified and the nonlinear model developed can be used to explain and predict the train actual DT well, and that the developed model can be applied to assist decision making at both the tactical and operational levels.
Modeling Actual Dwell Time for Rail Transit Using Data Analytics and Support Vector Regression
Jiang, Zhibin (author) / Gu, Jinjing (author) / Han, Yanzhao (author) / Fan, Wei (David) (author) / Chen, Jingjing (author)
2018-08-30
Article (Journal)
Electronic Resource
Unknown
Transit Assignment Model Incorporating Bus Dwell Time
| British Library Online Contents | 2013
Dwell Time Relationships for Light Rail Systems
| British Library Online Contents | 1992
| British Library Online Contents | 2016
Application of Dwell Time Functions in Transit Assignment Model
| British Library Online Contents | 2002