报告题目：Stochastic PDEs constrained shape/interface optimization

报 告 人：赵文举教授

报告时间：2019年6月5日 15:30-16:30  

报告地点：数学楼202室

报告摘要：

In this present, the stochastic shape/interface optimal control is computationally considered. The control strategy is to minimize the expectation of a tracking cost functional or energy dissipation constrained by a stochastic interface elliptic equation or stochastic Navier-Stokes equations. Stochastic shape variations are used to establish a decreasing sequence of admissible interfaces. The finite element method is used to discretize the state and adjoint systems and provides mesh moving direction. To reduce the computational complexity for uncertainty quantification, the sparse grid collocation method is applied to match the probability distribution for the relatively large scale optimization problems and stochastic sampling-based descent method is considered for the large-scale sampling optimization. The unrestricted/fixed volume/ fixed surface area constraints are further applied and considered in practical sense. Finally, the numerical results are provided to demonstrate the efficiency and effectiveness of our algorithms.

报告人简介：

   Wenju Zhao is postdoc in department of mathematics at Southern University of Science and Technology, Shenzhen. He was awarded Master of Science degree in Computational Mathematics Jilin University. He was awarded Doctor of Philosophy degree in Computational Science at Florida State University, USA.