<p>The multi-marginal Monge Kantorovich problem <b>(MKP)</b> has become a pivotal framework in probability theory, mathematical physics, and economics. Despite its significance, the complexity of its solution structure and uniqueness continues to pose difficulties. In this work, we examine the properties of optimal transport plans under the <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(m\)</EquationSource> <EquationSource Format="MATHML"><math> <mi>m</mi> </math></EquationSource> </InlineEquation>-twist condition, which ensures that any optimal plan is confined to the union of graphs of at most <InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(m\)</EquationSource> <EquationSource Format="MATHML"><math> <mi>m</mi> </math></EquationSource> </InlineEquation> transport maps, offering a precise characterization of solutions. Beyond this structural result, we explore alternative formulations of <b>(MKP)</b> problem, including its decomposition into two-marginal transport problems, which offer a decompositional approach. Specifically, we show that <b>(MKP)</b> can be rewritten as the center-marginal Monge Kantorovich problem <b>(</b><i>C</i><b>-MKP)</b> or adjacent-marginal Monge Kantorovich problem <b>(</b><i>A</i><b>-MKP)</b>, each of which retains structural properties from the classical two-marginal setting. Our results refine the understanding of optimality and uniqueness in <b>(MKP)</b>, bridging between two-marginal classical optimal transport theory and multi-marginal formulations. They have implications for applications in density functional theory, economics, and quantum transport by providing rigorous conditions for when solutions are concentrated on finitely many maps. By introducing new structural conditions techniques, this work advances our understanding of multi-marginal optimal transport.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Extremal Structures and Uniqueness of Multimarginal Optimal Transportation Via Two-Marginal Reduction

  • Mohammad Ali Ahmadpoor,
  • Abbas Moameni

摘要

The multi-marginal Monge Kantorovich problem (MKP) has become a pivotal framework in probability theory, mathematical physics, and economics. Despite its significance, the complexity of its solution structure and uniqueness continues to pose difficulties. In this work, we examine the properties of optimal transport plans under the \(m\) m -twist condition, which ensures that any optimal plan is confined to the union of graphs of at most \(m\) m transport maps, offering a precise characterization of solutions. Beyond this structural result, we explore alternative formulations of (MKP) problem, including its decomposition into two-marginal transport problems, which offer a decompositional approach. Specifically, we show that (MKP) can be rewritten as the center-marginal Monge Kantorovich problem (C-MKP) or adjacent-marginal Monge Kantorovich problem (A-MKP), each of which retains structural properties from the classical two-marginal setting. Our results refine the understanding of optimality and uniqueness in (MKP), bridging between two-marginal classical optimal transport theory and multi-marginal formulations. They have implications for applications in density functional theory, economics, and quantum transport by providing rigorous conditions for when solutions are concentrated on finitely many maps. By introducing new structural conditions techniques, this work advances our understanding of multi-marginal optimal transport.