Scalable Methods for Optimal Experimental Design and Optimal Control for Systems Governed by PDEs wi

来源 :第八届工业与应用数学国际大会 | 被引量 : 0次 | 上传用户：dl_smh

【摘要】

：

　　We formulate an A-optimal experimental design criterion for infinitedimensional nonlinear Bayesian inverse problems.Our method aims to minimize the average

【作者】

：

AlexanderianAlen;

【机构】

：

The Univ.of Texas at Austin

【出处】

：

第八届工业与应用数学国际大会

【发表日期】

：

2015年期

下载到本地 , 更方便阅读

下载此文赞助VIP

声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架

论文部分内容阅读

　　We formulate an A-optimal experimental design criterion for infinitedimensional nonlinear Bayesian inverse problems.Our method aims to minimize the average variance of a Gaussian approximation to posterior law of inversion parameters by solving a bi-level PDE-constrained optimization problem.I will also discuss risk-averse optimization under uncertainty with application to optimal control of PDEs with uncertain parameter fields.We use numerical results for a porous medium flow problem with uncertain permeability to illustrate the methods.

其他文献

A Numerical Far Field Boundary Condition for Anisotropic Laplace Operators

　　We propose a numerical far field boundary condition for anisotropic Laplace operator on arbitrary star-shaped domains.With a proper change of variable in th

会议

Some DG Methods for Singularly Perturbed Volterra Integro-Differential Equations

　　To our knowledge,there are much less works on the uniform convergence of numerical methods for singularly perturbed Volterra integrodifferential equations(V

会议

Stochastic Collocation Method for Solving PDEs with Random Coefficients

　　We develop a technique based on a polynomial chaos expansion for solving Darcy equation in stochastic porous media.For the input hydraulic conductivity rand

会议

Localization and Nonlinear Control for Small Rotary Unmanned Aerial Vehicles

　　Comparing with fixed wing unmanned aerial vehicles,the miniature rotary unmanned aerial vehicle has a simpler and compacter mechanical configuration,and is

会议

Fast Bayesian Optimal Experimental Design and Its Applications

　　We analyze Laplace method in the context of optimal Bayesian experimental design and extend this method from the classical scenario,where a single dominant

会议

Stability Issues in MC Integration of SDEs

　　One often encounters difficulties in Monte-Carlo simulations of complex stochastic processes.For example,sometimes hyperbolic components appear in a formal

会议

A Highly Scalable Communication-free Domain Decomposition BIE-WOS(Boundary Integral Equation-Walk on

　　A hybrid approach for solving Laplace equation in 3-D domains is presented.It bases on a local method for the Dirichlet-Neumann(DtN)mapping of a Laplace equ

会议

Optimal Sequential Experimental Design Using Dynamic Programming and Transport Maps

　　How to select a sequence of experiments that maximizes value of experimental data? We formulate this optimal sequential experimental design problem by maxim

会议

Computational Methods for Parameter Estimation and Optimal Experimental Design

　　Experimentalists face the dilemma of choosing between the accuracy and costs of an experiment.Optimization methods form the basic computational tools to add

会议

Bifurcation Diagrams of Positive Solutions of A Prescribed Mean Curvature Problem

　　We study global bifurcation diagrams and exact multiplicity of positive solutions for the one-dimensional prescribed mean curvature problem arising in Micro

会议

Scalable Methods for Optimal Experimental Design and Optimal Control for Systems Governed by PDEs wi

其他学术论文