Exact solutions and doubly efficient approximations of jump-diffusion itô equations (Q4223643)

The author presents results on convergence of equidistant, but jump-adapted approximations of Itô stochastic differential equations with white noise and adapted Poisson jumps under the common assumptions of uniform Lipschitz continuous coefficients. Numerical methods of first and second order of accuracy both in the mean square and weak sense are systematically derived by the help of generator and Taylor expansions. Their convergence is proven by an extension of \textit{G. N. Milstein}'s central techniques [``Numerical integration of stochastic differential equations'' (1994; Zbl 0810.65144)] to the case of jump diffusions. For numerical testing, the nonlinear, one-dimensional equation of the extended CIR model with multiplicative Poisson jumps (with some application to interest rate modelling in mathematical finance) is analytically solved and numerically approximated for comparison purposes.