Optimization Of Fuzzy Inference System Using Modified Genetic Algorithm

Fuzzy inference systems have long been used for modeling non-linear, uncertain and ambiguous systems in variety of engineering applications. However, these systems suffer from lower accuracy of forecasting due to their intuitive and subjective designs. A modified genetic algorithm is used in this paper to enhance the forecasting accuracy of fuzzy systems. Optimization of system design and automatic generation of fuzzy rule-base have been discussed. A chaotic time series, obtained by solving the Mackey-Glass differential equation, has been used to evaluate performance of the optimized fuzzy system. Simulation results clearly show improvement in the accuracy of fuzzy forecasting.