As with the advances in location-based technology that emerged from e-commerce and wireless communication, individual vehicular information is becoming available instantaneously and could be used for traffic signal control. A new paradigm of a signal control algorithm might be needed to accommodate such abundant location-based information. This preliminary study investigates the benefit of adaptive signal control under such conditions. A simple event-based simulation program along with a genetic algorithm-based signal optimization program developed in this study evaluates various scenarios including both demand and market penetration levels. A test bed used in the study consists of two one-way streets with single-lane approaches. This study provides a ceiling benchmark of adaptive signal control under perfect knowledge in vehicle arrivals. The results indicate that adaptive signal control outperforms pretimed signal control for all cases considered, but the marginal benefit of adaptive signal control increases up to a certain volume level and then decreases. The results also indicate that the use of imperfect knowledge improves performance of adaptive signal timing plans over pretimed timing plans.
Abstract