Rapid transit line performance analysis: A stochastic approach

Capacity, one of the basic characteristics of rapid transit line performance, has been usually defined as the maximum number of trains that can pass a point during one hour of operation. This number has been determined by the station with the longest dwell time, assuming deterministic operation. This research utilized systems approach that included operation and interactions among many trains along an entire line. Simulation model was used to perform an analysis of the stochastic nature of passenger boarding/alighting process at stations and the resulting random dwell time, headways, delays and load factors along the line. The line performance output was then analyzed for its sensitivity to the changes of various design, operational and human factors relevant to the rapid transit line operation. Research results indicated that the randomness of dwell times at stations has a strong impact on rapid transit line performance, and, particularly, on the line capability to operate certain maximum number of trains per hour at a given level-of-service, defined by the percent of trains affected by delays and load factor variability. The line operation showed significant sensitivity to various design and operational dwell time determinants, such as the position of platform exit(s)/entrance(s), number of doors per train, and line demand pattern defined by passenger volumes at all stations along the line. Also, from the analysis of several operational scenarios, such as operation of alternate long and short train consists, and operation with short-term fluctuations of the demand (peak-within-peak), it was found that the negative effects of the randomness in operation were intensified under these operational conditions. The most effective way to control the randomness in operation and, particularly, to improve the line maximum performance, has been found to be headway control that includes not only holding of trains that are ahead of the schedule (or follow after short time interval), but also, acceleration of trains that are behind the schedule (or follow after long time interval)