Existence and uniqueness of the solutions and convergence of semi-implicit Euler methods for stochastic pantograph equations (Q853991)

Consider the \(d\)-dimensional nonlinear stochastic pantograph equations of the Itô-type \[ dX(t) = f(t,X(t),X(qt)) dt + g(t,X(t),X(qt)) dB(t), \quad X(0)=X_0, \] with delay parameter \(0 < q < 1\) and time \(t \in [0,T]\), driven by an \(m\)-dimensional Brownian motion \(B\). First, the authors verify sufficient conditions for existence and uniqueness of the solutions of these equations such as the local Lipschitz and the linear growth conditions. For this purpose, they apply standard contraction mapping principles on Banach spaces \(L^2\) with appropriate standard metric. Second, under the usual Lipschitz and the linear growth conditions, it is proved that the drift-implicit Euler methods (in fact, stochastic \(\theta\)-methods with \(\theta \in [0,1]\)) for these equations converge in strong mean square sense with rate \(\frac{1}{2}\) along equidistant partitions of finite time-intervals \([0,T]\). Mean square stability of this numerical method is investigated under certain choices of step size \(h<1\) and Lipschitz-constants \(K\) with \(6 K h < 1\) as well. The leading error coefficient is estimated in terms of system parameters, too. Finally, some numerical illustrations for two nonlinear examples with \(\theta = 0\) and \(\theta = 0.5\) are presented.