This book is devoted to mean-square and weak approximations of solutions of shastic differential equations (SDE). These approximations represent two fundamental aspects in the contemporary theory of SDE. Firstly, the construction of numerical methods for such systems is important as the solutions provided serve as characteristics for a number of mathematical physics problems. Secondly, the employment of probability representations together with a Monte Carlo method allows us to reduce the solution of complex multidimensional problems of mathematical physics to the integration of shastic equations.
Along with a general theory of numerical integrations of such systems, both in the mean-square and the weak sense, a number of concrete and sufficiently constructive numerical schemes are considered. Various applications and particularly the approximate calculation of Wiener integrals are also dealt with.
This book is of interest to graduate students in the mathematical, physical and engineering sciences, and to specialists whose work involves differential equations, mathematical physics, numerical mathematics, the theory of random processes, estimation and control theory.
Table of ContentsIntroduction. 1: Mean-square approximation of solutions of systems of stochastic differential equations. 1. Theorem on the order of convergence (theorem on the relation between approximation on a finite interval and one-step approximation). 2. Methods based on an analog of Taylor expansion of the solution. 3. Explicit and implicit methods of order 3/2 for systems with additive noises. 4. Optimal integration methods for linear systems with additive noises. 5. A strengthening of the main convergence theorem. 2: Modeling of Itô integrals. 6. Modeling Itô integrals depending on a single noise. 7. Modeling Itô integrals depending on several noises. 3: Weak approximation of solutions of systems of stochastic differential equations. 8. One-step approximation. 9. The main theorem on convergence of weak approximations and methods of order of accuracy two. 10. A method of order of accuracy three for systems with additive noises. 11. An implicit method. 12. Reducing the error of the Monte Carlo method. 4: Application of the numerical integration of stochastic equations for the Monte Carlo computation of Wiener integrals. 13. Methods of order of accuracy two for computing Wiener integrals of functionals of integral type. 14. Methods of order of accuracy four for computing Wiener integrals of functionals of exponential type. Bibliography. Index.