Emergence of effective fuzzy rules for controlling mobile robots using DNA coding method

Abstract

Studies the discovery of fuzzy rules for fuzzy inference systems with multiple inputs. Recently, many research studies on fuzzy inference systems using genetic algorithms have been done. However, it is difficult to find effective rules if the system has many inputs and outputs. We have proposed a new coding method based on biological DNA and a mechanism of development from artificial DNA. We call this the "DNA coding method". This method has redundancy and overlapping genes. This paper applies the DNA coding method to the discovery of fuzzy rules. Selection of input variables and tuning of membership functions are carried out. Effective fuzzy rules for mobile robots are found through chasing and avoiding operations. An application of virus and enzyme operators to the artificial DNA is also presented in this paper.