A GA-based fuzzy traffic simulation for crossroad management

Conventional fuzzy traffic controllers use membership functions given by human operators. However, this approach does not guarantee the optimal solution to design fuzzy control systems. To find near optimal fuzzy membership functions, we perform traffic simulations by using genetic algorithms. However, it is not easy in traffic control to define a fitness function as a mathematical expression. Thus, we choose a simulation approach such that the fitness value of a solution is determined by using a performance measure obtained during traffic simulation. The experimental results show that the proposed fuzzy logic controller outperforms a conventional fuzzy controller in terms of the average delay and the average cost