An improved response surface methodology algorithm with an application to traffic signal optimization for urban networks

Illustrates the use of the simulation-optimization technique of response surface methodology (RSM) in traffic signal optimization of urban networks. It also quantifies the gains of using the common random number (CRN) variance reduction strategy in such an optimization procedure. An enhanced RSM algorithm which employs conjugate gradient search techniques and successive second-order models is presented instead of the conventional approach. An illustrative example using an urban traffic network exhibits the superiority of using the CRN strategy over direct simulation in performing traffic signal optimization. The relative performance of the two strategies is quantified with computational results using the total network-wide delay as the measure of effectiveness.