基于线性同余算法的格点方法在亚式期权定价中的应用

The linear-congruential-algorithm based lattice method in Asian Option Pricing

在拟蒙特卡罗方法中,低偏差序列性能的好坏直接决定拟蒙特卡罗估计的有效性,一般常用的拟蒙特卡罗估计使用的是基于(t,m,s)网的低偏序列,然而由这种方法产生的低偏序列在高维时有严重的聚丛现象,对高维时的估计精度有较大影响.使用基于线性同余算法的格点方法估计高维亚式期权的价格,比较了两种方法的计算精度和计算时间,表明格点方法在高维有很好的效果.

In Quasi-Monte Carlo method, the function of low-discrepancy sequences determines the efficiency of the estimation and may result in severe clumping phenomena in higher dimension situation and effect the precision of the estimation cousiderably. This paper uses the lattice metmod which is based on linear congruential algorithm to estimate the price of asian option. Comparison of these two methods in precision and time consumed, shows that lattice method works better under high dimension circumstance.