A chaos search immune algorithm with its application to neuro-fuzzy controller design

In this paper, a chaos search immune algorithm (CSIA) is proposed by integrating the chaos optimization algorithm and the clonal selection algorithm. First, optimization variables are expressed by chaotic variables through solution space transformation. Then, taking advantages of the ergodic and stochastic properties of chaotic variables, a chaos search is performed in the neighbourhoods of high affinity antibodies to exploit local solution space, and the motion of the chaotic variables in their ergodic space is used to explore the whole solution space. Furthermore, a generalized radial basis function neuro-fuzzy controller (GRBFNFC) is constructed and designed automatically by the proposed CSIA. Application of the CSIA-designed GRBFNFC to real-time control of an inverted pendulum system is discussed. Experimental results demonstrate that the designed GRBFNFC has very satisfactory performance.