Applying Genetic Algorithms for Traffic Light Control

1. Background

Owing to rapid urbanization and population growth, there is a pressing need for an optimized traffic control system that minimizes delays. The crux of this problem lies in the challenge of harmonizing behavior between traffic signals to achieve the common goal of reduced network delay. The efficient red and green signal times are determined by a known fitness function of a genetic algorithm. The major part in this work is concerned with applying genetic algorithms to find optimized scheduling solutions for efficient traffic flow. Genetic algorithms are introduced in the system to provide an intelligent green interval response based on the vehicular count.

2. Algorithm and Implementation

A simulator generated the fitness function, expressed as the signal settings, which was iteratively fed back to the genetic algorithm. Each approach made use of a green period and a red period. Chromosomes described the behavior of n traffic controllers. The first step was to compute optimal signal settings. The genetic search operators: Reproduction, Crossover and mutation were applied for the purpose of calculation of the fitness function. The analysis of data indicates that GA is best in finding near optimum.

3. Results

The graph denotes the fitness function of the offspring at each generation. At each iteration of the algorithm ie, the fitness is calculated and the bar graph is plotted. The bar graph illustrates the change in the fitness over time. We can see the fitness being optimized over the course of time.

4. Software Specifications

Implemented with MatLab and Python