variavel0=<b>Leandro dos Santos Coelho</b> - lscoelho@rla01.pucpr.br Pont. Univ. Católica - PR / Univ. Tuiuti do Paraná
<b>Abstract.</b> This paper presents the design of a Takagi-Sugeno-Kang fuzzy system based on simplex method for optimization for the antecedent part and least mean square for consequent part design. The fuzzy system is evaluated in two case studies: (i) prediction of CO2 concentration of Box and Jenkins gas furnace data, and (i) prediction of chaotic system wtih maps involving non-differentiable functions called Lozi map. Simulations present the estimation and validation procedures of dynamic models obtained by fuzzy system. The simulation results indicate that the fuzzy system of Takagi-Sugeno-Kang is attractive to applications of nonlinear identification, prediction of time series and design of advanced control algorithms.
<b>Keywords.</b> fuzzy system, nonlinear identification, dynamic systems, artificial intelligence, chaotic time series.