Using Evolutionary Programming to Control Metering Rates on Freeway Ramps

Abstract

A rule-based strategy is proposed for freeway ramp metering control. The rule-based controller determines ramp discharge rates based on upstream, downstream, and adjacent freeway occupancy levels. The use of a rule-base is advantageous because it can be implemented in real-time via a look-up table and because knowledge from traffic control engineers can be directly incorporated into the control strategy. Evolutionary programming is used to adapt rule thresholds so that this general procedure can be applied to any freeway configuration. A rule set is presented that operates on sensed occupancy rates. Evolutionary optimization of these sensor occupancy rate thresholds is described. The proposed method is then evaluated using INTRAS, a sophisticated traffic simulation program, on a uniform freeway segment. The results demonstrate that evolutionary programming is a viable approach to achieving ramp metering control strategies that can alleviate both recurrent and nonrecurrent freeway congestion. The presented study focuses on ramp meter control which, by itself, is effective only over a narrow regime of traffic flow conditions. Consideration is not given to ramp queues, surface street congestions, or traffic diversion.