This paper discusses a simulation effort to estimate the benefit of adaptive signal control when compared to pre-timed traffic signal timing plans. The study compared the performance of the two methods of signal control on a simplified, hypothetical intersection of two one-way roadways with single lane approaches. The modeled adaptive signal control system uses a genetic algorithm-based signal optimization technique to determine the most appropriate signal timing plan for current traffic conditions. Several scenarios of traffic flow variations over time, as well as a range of values for the accuracy of traffic detection used by the adaptive system were analyzed.

Comparing both signal control systems under conditions of perfect knowledge, the study found that adaptive control outperformed pre-timed signal plans across each of four traffic flow variation scenarios tested, resulting in 18 to 20 percent less delay. Under conditions of imperfect knowledge of traffic flow, representing varying levels of sensor deployment to monitor vehicles, the study found that adaptive signal control consistently outperformed pre-timed plans developed without knowledge of the traffic flows. The marginal benefit of adaptive signal control declined as traffic volumes increased, likely due to the ability of adaptive signal control to accommodate the greater fluctuations of demand during low volume scenarios. The authors noted that this was an exploratory simulation of adaptive signal control to establish a theoretical ceiling benchmark of adaptive signal control under conditions of perfect knowledge. They further noted that the study could be improved with better simulation of actual traffic conditions including multilane approaches and interactions between vehicles.