Uncertainty propagation and quantification are of great interest in computational electromagnetic simulations. It is particularly relevant in the context of electromagnetic compatibility where high-dimensional random variables problems rules. However, solving multirandom variable systems is an exceedingly difficult task due to the notoriously challenging problem known as “the curse of dimensionality”. In this article, we propose to associate a domain decomposition method with a stochastic technique to face this problem. First, an original fully asynchronous time-domain domain decomposition approach is introduced and illustrated for a transmission line network. Second, its association with the stochastic collocation method is presented and validated for a stochastic configuration. Finally, we discuss the dramatic reduction in the computational cost for time-domain stochastic simulations allowed by this domain-decomposition and variable-separation strategy.

中文翻译：

---

具有成本效益的传输线时域随机仿真的域分解方法


不确定性传播和量化在计算电磁仿真中非常有趣。它在高维随机变量问题占主导地位的电磁兼容性背景下尤其重要。然而，由于众所周知的“维数灾难”这一极具挑战性的问题，解决多随机变量系统是一项极其困难的任务。在本文中，我们建议将域分解方法与随机技术相关联来解决这个问题。首先，介绍并说明了一种针对传输线网络的原始完全异步时域分解方法。其次，提出了它与随机配置方法的关联，并针对随机配置进行了验证。最后，我们讨论了这种域分解和变量分离策略所允许的时域随机模拟计算成本的显着降低。