Dynamic signal optimization for traffic networks

An adaptive real-time applicable optimization method for traffic signal control in road networks is presented. The approach is based on a suitable macroscopic traffic model that is sufficiently accurate and computationally efficient. A model predictive control strategy is utilized to optimize traffic signaling throughout the traffic network. Fluctuating demands and changing road capacity are taken into account at every control interval. The control strategy, i. e. the choice of optimization goals, is flexible can be set on-line. A combination of optimization goals can be assigned to each link. The introduced concept is validated via simulation on a generic traffic network with the optimization goal to minimize average jam length. Comparison with a classic signal optimization that yields a static signal setting shows that the proposed method is capable of clearing traffic jams and thereby increases overall traffic throughput.