A recursive sparse grid collocation method for differential equations with white noise (Q6486748)

A sparse grid collocation method is an important technique for the numerical approach of the expectation of the solutions of stochastic equations combined with the Euler scheme. In this paper, it is shown that this method can have satisfactory accuracy for small noise and small integration time, but does not converge if the time step decreases or the grid level increases. To overcome this flaw, the authors combine this method with a recursive procedure to allow longer time treatment. Numerical examples are illustrated.