English

From Nonsmooth Minima to Smooth Branches via Heat Kernel Regularization

Optimization and Control 2026-04-08 v1

Abstract

Many optimization problems in science and engineering involve objective functions that are nonsmooth at their minimizers. A common strategy is to trace a branch of minimizers of a regularized objective as the smoothing scale tends to zero; however, for nonsmooth functions, it is generally unclear whether such a branch can be continued and whether the associated continuation equation remains locally solvable. We study heat-kernel regularization and the resulting continuation equation along a local minimizing branch connected to a minimizer of the original objective. Under a global growth condition and a local leading-order description of the form xa|x|^a with 1a21 \le a \le 2, we first show that the regularized objective admits global minimizers and that any such minimizing branch localizes at the natural heat scale O(t)O(\sqrt{t}). We then prove that the asymptotic behavior of the regularized Hessian is determined by the local profile of the original objective: it remains uniformly positive definite in the quadratic case a=2a=2, while in the subquadratic regime 1a<21 \le a < 2 its smallest eigenvalue grows at the controlled rate t(a2)/2t^{(a-2)/2}. Consequently, the regularized Hessian remains asymptotically nondegenerate for all sufficiently small t>0t>0, and the continuation equation remains locally solvable, even when the original objective does not admit a classical Hessian at the minimizer. Our results provide a rigorous second-order framework for continuation-based analysis in nonsmooth optimization by showing how heat regularization restores nondegeneracy near singular minimizers.

Keywords

Cite

@article{arxiv.2604.05372,
  title  = {From Nonsmooth Minima to Smooth Branches via Heat Kernel Regularization},
  author = {Hyeontae Jo},
  journal= {arXiv preprint arXiv:2604.05372},
  year   = {2026}
}
R2 v1 2026-07-01T11:56:32.834Z