Approximation and convergence of solutions to semilinear stochastic evolution equations with jumps (Q393062)

The aim of the paper is to prove the continuous dependence on initial data and the coefficients of the mild solution to the semilinear stochastic evolution equation NEWLINE\[NEWLINEdu(t) +Au(t)dt + f(u(t))dt = B(u(t-))dM(t),\quad u(0)=u_0,NEWLINE\]NEWLINE on a real separable Hilbert space \(H\) in which \(A:D(A)\subseteq H\to H\) is a linear maximal quasi-monotone operator, \(M\) is an \(H\)-valued discontinuous square-integrable martingale, and \(f\) and \(B\) are Lipschitz with linear growth condition. Sufficient conditions for the mild solution to be continuous with respect to \((u_0, A, f, B)\) are given. Since the factorization method is no longer applicable, the authors appeal to some techniques from the theory of nonlinear maximal monotone operators. The main part of the paper is to investigate the continuity of the stochastic convolutions with respect to the operator \(A\) by means of introducing the Yosida approximations and a limiting process. Here, the nonlinear version of the Trotter-Kato theorem is used.