Iterative surrogate model optimization (ISMO): An active learning algorithm for PDE constrained optimization with deep neural networks

Kjetil Olsen Lye

doi:10.1016/j.cma.2020.113575

Iterative surrogate model optimization (ISMO): An active learning algorithm for PDE constrained optimization with deep neural networks

2021-02-01

47

OA

AI

Praveen Chandrashekar

DOI：10.1016/j.cma.2020.113575

Original

Save

View PDF

Abstract

En 中文

We present a novel active learning algorithm, termed as iterative surrogate model optimization (ISMO), for robust and efficient numerical approximation of PDE constrained optimization problems. This algorithm is based on deep neural networks and its key feature is the iterative selection of training data through a feedback loop between deep neural networks and any underlying standard optimization algorithm. Numerical examples for optimal control, parameter identification and shape optimization problems for PDEs are provided to demonstrate that ISMO significantly outperforms a standard deep neural network based surrogate optimization algorithm as well as standard optimization algorithms. (C) 2020 The Author(s). Published by Elsevier B.V.

Keywords:

CFD

Deep learning

Optimization

Neural networks

Iterative surrogate model optimization (ISMO): An active learning algorithm for PDE constrained optimization with deep neural networks

Abstract

AI Summary

Journal

Organization