Problèmes de transport multi-marginal de Monge-Kantorovich : une caractérisation des solutions (Q479933)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Problèmes de transport multi-marginal de Monge-Kantorovich : une caractérisation des solutions
scientific article

    Statements

    Problèmes de transport multi-marginal de Monge-Kantorovich : une caractérisation des solutions (English)
    0 references
    0 references
    5 December 2014
    0 references
    The author studies the Monge-Kantorovich transport problem for Borel probability measures \(\mu_1,\mu_2,\dots, \mu_n\) on smooth manifolds \(X_1,X_2,\dots, X_n\) in the case where the cost function \(c:X_1\times X_2\times \dots \times X_n\to \mathbb{R}\) is bounded and continuous. Denoting by \(\Pi(\mu_1,\dots ,\mu_n)\) the set of Borel probability measures on \( X_1\times X_2\times \dots \times X_n\), it follows that \(X_i\) has marginal \(\mu_i\) for each \(i\in\{1,2,\dots ,n\}\). The transport cost associated with a transport plan \(\pi\in \Pi(\mu_1,\dots,\mu_n)\) is given by \(l_c(\pi)=\int_{X_1\times \dots \times X_n}c(x_1,\dots, x_n)d\pi.\) The author considers the Monge-Kantorovich transport problem \(\inf\{l_c(\pi); \pi\in \Pi (\mu_1,\dots ,\mu_n)\}\). One studies the support of optimal plans, where the cost function does not satisfy the classical twist condition in the two marginal problem as well as in the multi-marginal case when twistedness is limited to certain subsets. The main result of the paper is Theorem 1.3. Assume that the cost function \(c\) satisfies the \(m\)-twist condition on \(c\)-splitting sets and \(\mu_1\) is non-atomic. Then for each optimal plan \(\gamma\) of (MK) with \(Supp(\gamma)\subset D_1(c)\), there exists \(k\leq m\), a sequence of non-negative measurable real functions \(\{\alpha_i\}_{i=1}^{m}\) on \(X_1\), and Borel measurable maps \(G_1,\dots ,G_k:X_1\to X_2\times \dots \times X_n\) such that \(\gamma = \sum_{i=1}^{k}\alpha_i(Id\times G_i)_\sharp\mu\), where \(\sum_{i=1}^k\alpha_i(x)=1\) for \(\mu_1\)-a.e. \(x\in X_1\).
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    Monge-Kantorovich transport problem
    0 references
    optimal transport plan
    0 references
    0 references
    0 references