English

Variational representations for N-cyclically monotone vector fields

Optimization and Control 2016-01-20 v3

Abstract

Given a convex bounded domain Ω\Omega in Rd{{\mathbb{R}}}^{d} and an integer N2N\geq 2, we associate to any jointly NN-monotone (N1)(N-1)-tuplet (u1,u2,...,uN1)(u_1, u_2,..., u_{N-1}) of vector fields from % \Omega into Rd\mathbb{R}^{d}, a Hamiltonian HH on Rd×Rd...×Rd{\mathbb{R}}^{d} \times {\mathbb{R}}^{d} ... \times {\mathbb{R}}^{d}, that is concave in the first variable, jointly convex in the last (N1)(N-1) variables such that for almost all % x\in \Omega, \hbox{(u1(x),u2(x),...,uN1(x))=2,...,NH(x,x,...,x)(u_1(x), u_2(x),..., u_{N-1}(x))= \nabla_{2,...,N} H(x,x,...,x). Moreover, HH is NN-sub-antisymmetric, meaning that \sum% \limits_{i=0}^{N-1}H(\sigma ^{i}(\mathbf{x}))\leq 0 for all x\mathbf{x}% =(x_{1},...,x_{N})\in \Omega ^{N}, σ\sigma being the cyclic permutation on Rd{\mathbb{R}}^{d} defined by σ(x1,x2,...,xN)=(x2,x3,...,xN,x1)\sigma (x_{1},x_2,...,x_{N})=(x_{2},x_{3},...,x_{N},x_{1}). Furthermore, HH is NN% -antisymmetric in a sense to be defined below. This can be seen as an extension of a theorem of E. Krauss, which associates to any monotone operator, a concave-convex antisymmetric saddle function. We also give various variational characterizations of vector fields that are almost everywhere NN-monotone, showing that they are dual to the class of measure preserving NN-involutions on Ω\Omega.

Keywords

Cite

@article{arxiv.1207.2408,
  title  = {Variational representations for N-cyclically monotone vector fields},
  author = {Alfred Galichon and Nassif Ghoussoub},
  journal= {arXiv preprint arXiv:1207.2408},
  year   = {2016}
}
R2 v1 2026-06-21T21:33:29.617Z