Random matching in 2D with exponent 2 for densities defined on unbounded sets
Preprint
Abstract
We consider here the Random Euclidean Matching with exponent 2 for distributions defined on unbounded sets in the plane. The case of the exponent 2 on the plane has been particularly studied because, in the case of smooth and positive distributions defined on a compact set, it is possible to exactly determine the leading behavior of the cost. This was possible starting from a conjecture by Caracciolo et al., proved by Ambrosio et al., which allows to pose the problem in terms of PDE. The case of densities defined throughout the plane presents further difficulties. In particular, here we consider the case of the Maxwellian and of the Gaussian distribution. For the Gaussian, estimates by Talagrand and Ledoux determine the leading behavior of the average of the cost but for a multiplicative constant. Here we determine exactly the leading behavior.Keywords
All Related Versions
This publication has 18 references indexed in Scilit:
- Random Assignment Problems on 2d ManifoldsJournal of Statistical Physics, 2021
- Euclidean Random Matching in 2D for Non-constant DensitiesJournal of Statistical Physics, 2020
- On Optimal Matching of Gaussian SamplesJournal of Mathematical Sciences, 2019
- Anomalous Scaling of the Optimal Cost in the One-Dimensional Random Assignment ProblemJournal of Statistical Physics, 2018
- A PDE approach to a 2-dimensional matching problemProbability Theory and Related Fields, 2018
- Remarks on the Monge–Kantorovich problem in the discrete settingComptes Rendus Mathematique, 2018
- Combinatorial Optimization Over Two Random Point Sets2013
- A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problemNumerische Mathematik, 2000
- Transportation cost for Gaussian and other product measuresGeometric and Functional Analysis, 1996
- Asymptotics for transportation cost in high dimensionsJournal of Theoretical Probability, 1995