A Derivative-Free Milstein Type Approximation Method for SPDEs covering the Non-Commutative Noise case
Higher order schemes for stochastic partial differential equations that do not possess commutative noise require the simulation of iterated stochastic integrals. In this work, we propose a derivative-free Milstein type scheme to approximate the mild solution of stochastic partial differential equations that need not to fulfill a commutativity condition for the noise term and which can flexibly be combined with some approximation method for the involved iterated integrals. Recently, the authors introduced two algorithms to simulate such iterated stochastic integrals; these clear the way for the implementation of the proposed higher order scheme. We prove the mean-square convergence of the introduced derivative-free Milstein type scheme which attains the same order as the original Milstein scheme. The original scheme, however, is definitely outperformed when the computational cost is taken into account additionally, that is, in terms of the effective order of convergence. We derive the effective order of convergence for the derivative-free Milstein type scheme analytically in the case that one of the recently proposed algorithms for the approximation of the iterated stochastic integrals is applied. Compared to the exponential Euler scheme and the original Milstein scheme, the proposed derivative-free Milstein type scheme possesses at least the same and in most cases even a higher effective order of convergence depending on the particular SPDE under consideration. These analytical results are illustrated and confirmed with numerical simulations.