Abstract
On the urban railway network system in Tokyo metropolitan area, the number of operating trains is set close to the capacity of the line at peak times so that the standard speed of the trains is markedly decreased and recovery takes a long time in the case of an incident that disturbs the operation. Therefore, we observed knock-on train delay and time required for alighting and boarding passengers which is one of the delay factors of a train.
This paper presents knock-on train delay estimation method using multi-agent model. The proposed models reflect the constraints of ATC signaling system, train operation rules and passenger behavior rules.
