Distributed management of Personal Rapid Transit (PRT) vehicles under unusual transport conditions

The paper presents a flexibility of management of vehicles in Personal Rapid Transit (PRT) network. The algorithm used for delivering empty vehicles for waiting passengers is based on multiparameter analysis. Due to its distributed construction, the algorithm has a horizon parameter, which specifies the maximum distance between stations the communications is performed. Every decision is made basing on an information about situation (number of vehicles standing at a station, number of vehicles travelling to a station, number of passengers waiting) sent between stations, without any central data base containing traffic conditions. The simulation of the traffic in random case (typical) and in unusual case of delivering people to a social event occurring at single place is presented. It is shown that simple manipulation with horizon parameter allows to adapt the network to extremely uneven demand and destination choice.