In this work numerical methods for the treatment of random ordinary differential equations are presented. The focus is on approaches, which are based on the generalized Wiener expansions. Especially the stochastic Galerkin method in connection with the multi-element approach by Wan and Karniadakis is discussed. By the application of this method, the random ordinary differential equation becomes a system of deterministic ordinary differential equations for the coefficients of the generalized Wiener expansion of the solution. The resulting system can be solved by Runge-Kutta methods. The convergence and stability properties of this method are discussed. Based on the theoretical results, an algorithm for the solution of random ordinary differential equations is presented. It incorporates a new error estimator, which combines the known error estimator for the time integration with the adaptive choice of elements in the parameter space of the uncertain data. Numerical results to validate the algorithm close the thesis.
«In this work numerical methods for the treatment of random ordinary differential equations are presented. The focus is on approaches, which are based on the generalized Wiener expansions. Especially the stochastic Galerkin method in connection with the multi-element approach by Wan and Karniadakis is discussed. By the application of this method, the random ordinary differential equation becomes a system of deterministic ordinary differential equations for the coefficients of the generalized Wien...
»