中文摘要：

为了克服极小概率事件发生概率估计的困难，提出了把重要抽样技术发展到外汇期权组合非线性VaR模型中，估计出组合损失概率。为了进一步达到减少模拟估计误差目的，在重要抽样技术基础上使用分层抽样技术，进行更有效的MonteCarlo模拟。数值结果表明，重要抽样技术算法比常用MonteCarlo模拟法的计算效率更有效；而重要抽样技术和分层抽样技术相结合算法比重要抽样技术算法更有效地减少模拟所要估计的组合损失概率的方差，有着更高的计算效率。

英文摘要：

To overcome the difficulty in estimating low probability, the importance samping technique is developed into non-linear VaR model of FX option portfolio, and the loss probability of portfolio is estimated precisely. Moreover, the algorithm combining importance samping technique with stratified samping technique is used to carry though more efficient Monte Carlo simulation. The simulation result shows the algorithm using importance samping technique has much more effectiveness of computational efficiency than the standard Monte Carlo simulation. And the algorithm combining importance samping with stratified samping has much more effectiveness of computational efficiency than the algorithm using importance samping technique, and can lead to larger variance reductions when estimating the loss probability of portfolio.