English

Trade-off Invariance Principle for minimizers of regularized functionals

Optimization and Control 2025-01-28 v4

Abstract

In this paper, we consider functionals of the form Hα(u)=F(u)+αG(u)H_\alpha(u)=F(u)+\alpha G(u) with α[0,+)\alpha\in[0,+\infty), where uu varies in a set UU\neq\emptyset (without further structure). We first revisit a result stating that, excluding at most countably many values of α\alpha, we have infHαG=supHαG\inf_{H_\alpha^\star}G= \sup_{H_\alpha^\star}G, where Hα:=argminUHαH_\alpha^\star := \arg\min_UH_\alpha, which is assumed to be non-empty. Then, we prove a stronger result that concerns the invariance of the limiting value of the functional GG along minimizing sequences for HαH_\alpha, which extends the above Principle to the case Hα=H_\alpha^\star= \emptyset. Moreover, we show to what extent these findings generalize to multi-regularized functionals and -- in the presence of an underlying differentiable structure -- to critical points. Finally, the main result implies an unexpected consequence for functionals regularized with uniformly convex norms: excluding again at most countably many values of α\alpha, it turns out that for a minimizing sequence, convergence to a minimizer in the weak or strong sense is equivalent.

Keywords

Cite

@article{arxiv.2411.11639,
  title  = {Trade-off Invariance Principle for minimizers of regularized functionals},
  author = {Massimo Fornasier and Jona Klemenc and Alessandro Scagliotti},
  journal= {arXiv preprint arXiv:2411.11639},
  year   = {2025}
}

Comments

20 pages, extension to multi-regularization and to critical points, differentiability of the value function

R2 v1 2026-06-28T20:03:38.731Z