A class of balanced stochastic Runge-Kutta methods for stiff SDE systems (Q2356068)

To strongly approximate the solution of stiff systems of Itô stochastic differential equations with multiplicative noise while reducing computational cost, a family of balanced stochastic Runge-Kutta methods that do not use derivatives is presented. It is proved that these balanced methods have strong convergence of order \(p\) when the Runge-Kutta methods used satisfy the conditions required for them to converge with strong order \(p\). For four specific balanced methods of this type, mean-square stability is analyzed and data on their accuracy for four test problems is summarized and discussed.