An efficient radial basis function generated finite difference meshfree scheme to price multi-dimensional PDEs in financial options (Q6049303)

In this paper, the author propose a new solver for pricing financial option at the presence of several assets. Briefly, the author's contributions are as follows: -- by using non-uniform distribution of the nodes, we construct and obtain numerical results which are efficient, while the size of the discretization system is as low as possible; -- it is intended to focus on solving the multi-dimensional PDE problem \[ \frac{\partial U}{\partial \tau}={\displaystyle \sum_{i=1}^M}(r-q_i)X_i\frac{\partial U}{\partial X_i}+\frac{1}{2}{\displaystyle \sum_{i,j=1}^M}\rho_{ij}\sigma_i\sigma_jX_iX_j\frac{\partial ^2U}{\partial X_i \partial X_j}-rU \] by encompassing the whole solving procedure into matrix notations employing sparse arrays so as to cope with the problem as efficiently as possible; -- since careless numerical computations may waste a good mathematical model, here the stability of the presented method is discussed and proved; -- it is proved and shown by way of illustration that the new solver is less sensitive to its shape parameter (the free non-zero parameter) and can yield in efficient approximations. Numerical tests from baskets having several risky assets are considered and solved in Section 5.