两类求解随机微分方程的数值方法-计算数学专业论文.docxVIP

Abs Abstract - - II - Abstract With the spreed of technology and science, both in the nature and the real life, the effect of uncertain factors is nonnegligible to the development state of things. Hence it is more precised to portray the motion law of the nature by using stochastic differential equations. But stochastic differential equations usually have strong the coupling and nonlinear property, it is difficult to calculate the expression of the solution, so we need to construct numerical methods to approximate the solutions of stochastic differential equations, and further simulate the behavior of solutions by using the mathematical tool and computer. In order to simulate the quantitative behavior of solutions more accurately, we require the constructed numerical methods to maintain the structure of the original stochastic system, such as symplectic structure, phase volumes and first integrals, etc. Firstly, our work introduces the research background of stochastic differential equations with its established purpose and significance. Then it summarizes the development status for numerical method of stochastic differential equations, both domestic and overseas. Secondly, we briefly introduces the required preliminary knowledge. Moreover, we construct two types of non-standard finite difference method which based on Euler-Maruyama method and Milstein method. By comparing the constructed non-standard finite difference method of one-step approximation and the deviation of the exact solution of the original equation, we obtain two types of non-standard finite difference method with the mean-square convergence order, and apply numerical examples to verify the theoretical results. Finally, for a special class of stochastic differential equations, this dissertation utilizes the stochastic Runge-Kutta methods to calculate the equations and deduces stochastic Runge-Kutta-Nystr?m methods. The numerical solution and the exact solution will be expanded as the stochastic Tayl