Rare event simulation method based on importance sampling technology

It usually takes long time to simulate rare event using traditional Monte Carlo method, while importance sampling techniques can effectively reduce the simulation time and improve simulation efficiency. A new implementation for importance sampling method to estimate rare event probability in simulation models was proposed. The classical exponential change of measure was adopted to construct the family of importance sampling distributions, and the optimal importance sampling distributions was obtained by minimizing the variance of importance sampling estimator. Numerical experiments have been conducted and the results indicate that the method can effectively estimate the rare event probabilities.