An optimal matching problem
Optimization and Control
2007-05-23 v1 Analysis of PDEs
Abstract
Given two measured spaces (X,dx), (Y,dy) and a third space Z, given two functions u(x,z) and v(x,z), we study the problem of finding two maps s from X to Z and t from Y to Z such that the images s(dx) and t(dy) coincide, and the integral of u(x,s(x))dx+v(y,t(y))dy is maximal. We give condition on u and v for which there is a unique solution.
Cite
@article{arxiv.math/0308206,
title = {An optimal matching problem},
author = {Ivar Ekeland},
journal= {arXiv preprint arXiv:math/0308206},
year = {2007}
}