Numerical Solutions for Optimal Control under SPDE Constraints
Abstract
The focus of this project is about numerical solutions of optimal control problems using the analysis of variance (ANOVA) analysis. The impact of random parameter dependent boundary conditions on the solutions of a class of nonlinear partial differential equations (PDEs) is considered. Because the boundary conditions are random field, the PDE becomes stochastic PDE. The concepts of effective dimensions are used to determine the accuracy of the ANOVA expansions. Demonstrations are given to show that whenever truncated ANOVA expansions of functionals provide accurate approximations, optimizers found through a simple surrogate optimization strategy are also relatively accurate.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2008
- Accession Number
- ADA479290
Entities
People
- Yanzhao Cao
Organizations
- Florida A&M University