Fuzzy robustness analysis based on importance sampling and neural network

This paper applies the importance sampling (IS) method and neural network (NN) to the fuzzy robustness analysis of uncertain control systems. The IS method is utilized to improve the sampling efficiency when the probability of fuzzy unacceptable performance is very small. The NN is used to predict the performance index requiring more computational time in each simulation experiment. The proposed approach can reduce the excessive computational cost generated from the standard Monte Carlo simulation (MCS) for dealing with the rare event case and the performance index requiring more computational time in the fuzzy robustness analysis. Finally, a numerical example is provided to demonstrate the effectiveness of the proposed method.