Two-norm discrepancy and convergence of the stochastic gradient method with application to shape optimization

TL;DR

This paper proves the convergence of the stochastic gradient method for a shape optimization problem involving random boundaries, addressing the two-norm discrepancy phenomenon with theoretical analysis and numerical validation.

Contribution

It extends stochastic gradient convergence theory to shape optimization problems with two-norm discrepancy, specifically for convex, smooth domains with random interior boundaries.

Findings

Convergence of stochastic gradient method is established for the shape optimization problem.

Theoretical analysis addresses the two-norm discrepancy phenomenon.

Numerical experiments validate the theoretical results.

Abstract

The present article is dedicated to proving convergence of the stochastic gradient method in case of random shape optimization problems. To that end, we consider Bernoulli's exterior free boundary problem with a random interior boundary. We recast this problem into a shape optimization problem by means of the minimization of the expected Dirichlet energy. By restricting ourselves to the class of convex, sufficiently smooth domains of bounded curvature, the shape optimization problem becomes strongly convex with respect to an appropriate norm. Since this norm is weaker than the differentiability norm, we are confronted with the so-called two-norm discrepancy, a well-known phenomenon from optimal control. We therefore need to adapt the convergence theory of the stochastic gradient method to this specific setting correspondingly. The theoretical findings are supported and validated by numerical experiments.