Mean-square stability properties of an adaptive time-stepping SDE solver (Q2496261)

Using the stochastic test equation \[ dX= \lambda X\,dt+ \mu X\, dW,\quad X(0)= X_0\neq 0 \] where \(W(t)\) is a standard Wiener process, the authors study the mean-square stability of an adaptive time-stepping Milstein method for approximating the solutions of stochastic differential equations. To determine the appropriate step sizes, two local error controls are devised, one based on the drift term and the other on the diffusion term of the test equation. Numerical results are presented and used to estimate the numerical stability boundary.