Bi-Level Optimization in a Transport Network

This paper applies a bi-level formalism for the optimal control of an urban transportation network. The well known store-and-forward model in traffic control is utilized in order to increase the control space of the optimization problem. Mainly, the store-and-forward models apply the split as a control argument, assuming the traffic light cycle as a constant parameter. The paper shows that by using a bi-level formalism the control problem can be defined within increased control space comprising both the split and the cycle. Both are found as optimal solutions of a bi-level optimization problem.