Linear mean-square stability analysis of weak order 2.0 semi-implicit Taylor schemes for scalar stochastic differential equations (Q1946154)

Mean square (MS) stability of semi-implicit weak order 2.0 Taylor schemes, called \((\theta,\beta)\) methods, for solving scalar Itô stochastic differential equations (SDE) of the form \[ dX_t= a(t,X_t)\,dt+ b(t, X_t)\,dW_t,\quad X_{t_0}= c, \] is studied. Relationships between the domain of stability of the SDE and that of \((\theta,\beta)\) methods approximating it are derived. Results of numerical experiments on the test SDE \(dX+t=\lambda X_t dt+\mu X_t dW_t\) are presented that show that the predicted stability behavior occurs. For \(0\leq\beta\leq 1\), a condition on the stepsize is established as necessary and sufficient for MS stability of the \(({1\over 2},\beta)\) scheme when the SDE being approximated is MS stable.