A transit passenger flow control model on KCR

Passengers moving from Hong Kong to Shenzhen exceed nearly 300 thousands in special days such as during Chinese new year holiday. Nearly half of these passengers use Kowloon-Canton Railway (KCR) to Chinese mainland through Lo Wu custom. Such heavy passenger traffic demand often exceeds the processing and holding capacity of Lo Wu station for many hours of a day. Therefore, passengers must be metered off at all stations dong KCR line through sales of limited number of tickets during each time period so as to hold passengers at Lo Wu departure hall within its holding capacity and passenger arrival rates are less than or equal to the processing rate at Lo Wu custom. This paper develops an optimal control strategy for the above passenger queuing and crowding problem. The control problem is formulated as successive linear programming problems for a sequence of time periods. Because the maximum processing rate at Lo Wu custom is fixed and passenger arrival rate at each station is inelastic, the total passenger waiting time cannot be reduced. Therefore, the objective is to minimize the differences among passenger waiting times at all KCR stations subject to capacity constraints. By successively solving a sequence of linear programming problems, optimal numbers of passengers to be released at each station during each period can be determined. The model is applied to the crowded KCR system with simulated data