A computational weighted finite difference method for American and barrier options in subdiffusive Black-Scholes model (Q2656030)

In this paper the authors derive the linear complementarity problem (LCP) system describing the fair price of an American option in subdiffusive Black-Scholes (B-S) model. The weighted scheme of the weighted finite difference (FD) method and the Longstaff-Schwartz method to solve the system numerically are considered. Some comparisons are also given. The following statement is precisely proved Theorem. The fair price an American put option in the subdiffusive (B-S) model is equal to \(\nu(z,t)\), where \(\nu(z,t)\) satisfies: \[ \begin{array}{l} x=\ln z\\ u(x,t)=\nu (e^x,T-t) \end{array} \] and \(u(x,t)\) is the solution of the system \[ \begin{array}{l} u(x,0)=\max (K-exp(x),0)\\ u(x,t)=\max (K-exp(x),0),\: \: t\in [0,T]\\ \begin{array}{l} c\\ o \end{array} D^{\alpha}u(x,t)-\frac{\sigma ^2}{2}\frac{\partial ^2 u(x,t)}{\partial x^2}-(r-\frac{\sigma ^2}{2})\frac{\partial u(x,t)}{\partial x}+ru(x,t)\ge 0,\: t\in (0,T)\\ (u(x,t)-\max (K-exp(x),0))(\begin{array}{l} c\\ o \end{array} D^{\alpha}u(x,t)-\frac{\sigma ^2}{2}\frac{\partial ^2 u(x,t)}{\partial x^2}-\\ (r-\frac{\sigma ^2}{2})\frac{\partial u(x,t)}{\partial x}+ru(x,t))=0,\: \: t\in (0,T)\\ {\displaystyle \lim_{x\to \infty}}u(x,t)=0\\ {\displaystyle \lim_{x\to -\infty}}u(x,t)=K. \end{array} \] The numerical techniques presented in this paper can successfully be repeated for other time fractional diffusion models with moving boundaries problems.