The problem of routing automatically-controlled and guided vehicles through a closed two-way personal rapid transit network is studied. Because of fluctuating passenger demands, a dynamical model is needed to describe the concentration of vehicles in the different network sections. Such a model is derived in this paper. The binary decision to be made at each divert junction is treated as a control variable in the system model. Using a suboptimal strategy to minimize the maximum congestion in the network sections served, the decision is made in a feedback manner based on the average concentrations in the sections. The model was used to study the transient behavior of flows and concentrations in a simulation of a specific closed two-way PRT network. Some results are presented for this example.
Abstract