Strong Convergence of the Tamed Euler-Maruyama Method for Stochastic Singular Initial Value Problems with Non-Globally Lipschitz Continuous Coefficients
Yan Li,Nan Deng,Wanrong Cao
DOI: https://doi.org/10.1016/j.apnum.2024.07.001
IF: 2.994
2024-01-01
Applied Numerical Mathematics
Abstract:In our previous works [1] and [2], we delved into numerical methods for solving stochastic singular initial value problems (SSIVPs) that involve coefficients satisfying the global Lipschitz condition. The paper addresses the limitations of our previous work by introducing an explicit method, called the tamed Euler-Maruyama method, for numerically solving SSIVPs with non-globally Lipschitz continuous coefficients, which is both easy-to-implement and exceptionally efficient. We prove the existence and uniqueness theorem and the boundedness of the moments of the solution to SSIVPs under the non-globally Lipschitz condition. Moreover, we provide a sharp analysis of the strong convergence of the proposed method, along with the uniform boundedness of numerical solutions. We also apply our results to the stochastic singular Ginzburg-Landau system and provide numerical simulations to illustrate our theoretical findings.