English

Entropic optimal transport need not select a zero-temperature limit

Optimization and Control 2026-07-18 v1 Probability

Abstract

We construct a compact metric space with an atomless probability measure and a bounded Lipschitz cost for which the entropic optimal-transport minimisers have no zero-temperature weak limit. More precisely, PεP_\varepsilon does not converge as ε0\varepsilon\downarrow0. In the example, every unregularised minimiser is singular with respect to μμ\mu\otimes\mu, so that the entropy on the optimal face is identically ++\infty. We describe the cluster set by Clust(Pε)={Pw:wW}, \operatorname{Clust}(P_\varepsilon)=\{P_w:w\in\mathcal W\}, where PwP_w is the mixture of the two zero-cost graph couplings with weight ww, and where W[0,1]\mathcal W\subset[0,1] is a non-degenerate compact interval. We then compute two explicit points w<w+w^-<w^+ in this interval. This shows that compactness, atomlessness, and Lipschitz regularity of the cost do not imply zero-temperature convergence. We also present a compactness theorem for the general problem. If CL1(μν)C\in L^1(\mu\otimes\nu) is continuous and bounded from below on Polish spaces, then the zero-temperature cluster set is a nonempty weakly compact connected subset of the optimal face. In the proof, we apply the cluster-point theorem of Bernton, Ghosal, and Nutz and the continuity of επε\varepsilon\mapsto\pi_\varepsilon. Finally, we give local and exterior first-order criteria for full convergence and cluster membership. We show that nonconvergence is possible, but only through a connected continuum of optimal plans.

Cite

@article{arxiv.2607.16881,
  title  = {Entropic optimal transport need not select a zero-temperature limit},
  author = {Maja Gwozdz},
  journal= {arXiv preprint arXiv:2607.16881},
  year   = {2026}
}