Self-generating rule-mapping fuzzy controller design using a genetic algorithm

A genetic algorithm (GA) based method is proposed to design a self-generating rule-mapping fuzzy controller. Its construction is based on the concept of a template rule base, as suggested by MacVicar & Whelan (see R.R. Yager et al., 1994). In the GA approach, an individual is constructed to represent a fuzzy controller. A short coded string is proposed such that, when associated with an individual, it can map a rule to a fuzzy controller structure, including the number of membership functions for each input variable, the shapes of the membership functions associated with each input variable and the index function. Then, a fitness function is proposed to guide the search procedure to select an appropriate fuzzy controller in order to satisfy the desired performance. Finally, the inverted pendulum control problem is utilised to illustrate the efficiency of the proposed method