A GREEDY KERNEL-BASED PARTITION OF UNITY METHOD FOR MULTI-ASSET AMERICAN OPTION PROBLEMS

In this paper we use a greedy algorithm on a kernel-based partition of unity method for the solution of multi-asset American option problems. This method can be overcome on a number of challenges for the kernel-based methods such as dense matrices and ill-conditioning. In the greedy algorithm the optimal trial points for interpolation obtained among a huge set of initial points and are used for kernel-based partition of unity method. This paper shows that selection nodes greedily yields the better conditioning and reduceing the computational cost for the option pricing problem.