Multivalued stochastic differential equations: Convergence of a numerical scheme (Q1410233)

The author studies the strong mean-square convergence of a numerical scheme for an \(R^d\)-multivalued stochastic differential equation \[ dX_t+A(X_t) dt \ni b(t,X_t) dt + \sigma(t, X_t) dW_t \] and obtains the rate of convergence \(O(\delta \log(1/\delta))^{1/2}\) when the diffusion coefficient is bounded. By introducing a discrete Skorokhod problem, he establishes \(L^p\)-estimates (\(p \geq 2\)) for the solutions. Some numerical results of the problem are presented.