Using genetic algorithms to design signal coordination for oversaturated networks

This article presents an algorithm to generate optimal (real-time) signal timings that distribute queues over a number of signalized intersections and over a number of cycles on any signalized intersection. A discrete-time signal-coordination model is formulated as a dynamic optimization problem and solved using Genetic Algorithms (GA). Signal timings for all intersections in the network during congested periods are decision variables and are represented in the individual GA candidate solutions. The algorithm is applied to a one-way arterial network with 20 signalized intersections. Depending on the traffic demand's variation and the position of critical signals, the algorithm intelligently generates optimal signal timing (offsets) along individual arterials. If critical signals are located at the exit points, the algorithm sets the optimal signal timing that protects them from becoming excessively loaded. If critical signals are located at the entry points, the algorithm ensures that queues are reduced or cleared before released platoons arrive at a downstream signal system. In this article, the simple genetic algorithm (SGA) with multiple epochs is used to solve the signal coordination problem. When a serial SGA is applied to solve traffic control problems, its performance in terms of computation time diminishes as the size of signal networks increases, or the duration of congestion lengthens. Master-slave SGA executed in a parallel computing machine is then used to reduce the execution time. We found that with a master-slave parallelism, SGA can be efficiently executed with significant speed-up, allowing the opportunity to implement the algorithm on real-time signal control systems.