使用似然比方法对期权敏感性进行估计时,分别采用标准的蒙特卡罗模拟、拟蒙特卡罗模拟和分层抽样下的蒙特卡罗模拟,比较三种方法在模拟结果的可靠性和耗时上的优劣,发现拟蒙特卡罗模拟和分层抽样下的蒙特卡罗模拟都优于标准的蒙特卡罗方法;拟蒙特卡罗模拟在模拟结果的可靠性上具有明显的优势,在模拟维数为一维时,拟蒙特卡罗模拟的模拟耗时并不明显优势,而在二维时,耗时明显比分层抽样下的蒙特卡罗模拟短。
With the likelihood ratio method, this paer estimates the sensitivities by standard Monte Carlo, Quasi - Monte Carlo and Monte Carlo with Stratified Sampling respectively and compared them in result's reliability and Time - consuming. Results show that Quasi - Monte Carlo and Monte Carlo with Stratified Sampling are both better than standard Monte Carlo, Quasi - Monte Carlo is superiority to Monte Carlo with Stratified Sampling in resuh's reliability. When the dimension is one, Quasi - Monte Carlo spend similar time with Monte Carlo with Stratified Sampling, but when it comes to two, the advantage of QMC is obvious.