An self-adaptive traffic signal fuzzy controller with improved GA-online-optimized parameters

Abstract

This thesis pushes forward a brand-new optimization method on the fuzzy controller realized by the improved GA algorithm at an isolated intersection, which is appropriate for binary code scheme with the constraint upon all three proportional factors. Provided taking on appropriate crossover and mutation operators, plus with classic roulette wheel selection, such an algorithm is able to solve problems like ill-generated individuals and the prematurity, then accelerates the convergence speed correspondingly. On the basis of the optimized membership functions and fuzzy rules, all three proportional factors are to be adjusted to the best simultaneously. The validity of this method has been proven by means of the MATLAB program simulation about the above-mentioned optimized fuzzy controller.