A Review of the Optimized Policies for Adaptive Control Strategy

Problem

Optimized Policies for Adaptive Control is a real-time, demand-responsive traffic signal timing optimization algorithm for individual intersections developed at the University of Lowell in the early '80s. It drops the concept of "cycle," instead constituting a sequence of switching decisions at fixed time intervals. The decision each time is whether to extend or terminate the current phase, and is constrained only by minimum and maximum defined phase lengths. It uses online and historical data to optimize the decisions. Although designed for individual intersections, the strategy can be used as a building block for decentralized networks of intersections.

This report reviews the design philosophy and underlying principles of OPAC, describes its development history and how it works, and provides a brief report on how well it performs.

Findings

The results of the field test show that OPAC strategies perform better overall than well-timed actuated signals, especially under heavy traffic conditions. However, it proved very sensitive to intersection discharge rates, so in order for it to perform well those must be accurately estimated.