摘要
考虑有限体积法求解Kou模型下美式跳扩散期权.基于线性有限元空间,构造了向后欧拉和Crank-Nicolson两种全离散有限体积格式,并采用简单高效的递推公式对偏微分积分方程中的积分项进行逼近.针对美式期权离散得到的线性互补问题(LCP),采用模超松弛迭代法(MSOR)进行求解,并证明了H_+离散矩阵下算法的收敛性.数值实验表明,所构造的方法是高效而稳健的.
Finite volume methods are developed for pricing American options under Kou jump-diffusion model. Based on a linear finite element space, both backward Euler and CrankNicolson full discrete finite volume schemes are constructed. For the approximation of the integral term in the partial integro-differential equation (PIDE), an easy-to-implement recursion formula is employed. Then we propose the modulus- based successive overrelaxation (MSOR) method for the resulting linear complementarity problems (LCPs). The H+ matrix property of the system matrix which guarantees the convergence of the MSOR method is analyzed. Numerical experiments confirm the efficiency and robustness of the proposed methods.
出处
《同济大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2016年第9期1458-1465,共8页
Journal of Tongji University:Natural Science
基金
国家自然科学基金(11271289)
中央高校基本科研业务费专项资金
云南省应用基础研究计划青年项目(2013FD045)
云南省教育厅科学研究基金项目(2015Y443)
关键词
有限体积法
Kou跳扩散期权模型
线性互补问题
模超松弛迭代法
finite volume method
Kou jump-diffusion option model
linear complementarity problem
modulus-based successive overrelaxation method