The unique minimal dual representation of a convex function (Q984820): Difference between revisions

This paper presents an interesting specification concerning the Fenchel--Moreau Theorem. The classical result states that, if \(X\) is a Banach space and \(f:X\to {\mathbb R}\cup \{+\infty\}\) is a lower semi--continuous and convex function, then \[ f(x)=\sup_{x^*\in X^*}[\langle x,x^*\rangle -f^*(x^*)], \] where \(f^*\) denotes the Fenchel transform of \(f.\) In the main result of this paper, under suitable assumptions on \(X,\) \(C\) and \(f,\) the smallest subset \({\mathcal M}_f\) of \(X^*\) is detected in order to recover the function \(f\) via its Fenchel transform, as in the equality above. This set \({\mathcal M}_f\) is related to the range of the subdifferential map \(x\mapsto \partial f(x),\) and it is defined as the closure in the weak\(^*\) topology of the set \[ {\mathcal N}_f=\{x^*\in X^*: x^*\in \partial f(x),\, x\in {C}_f\}, \] where \({C}_f\) is the set of points in \(C\) for which the subdifferential is a singleton. Indeed, the authors prove the following Theorem. Suppose that (i) \(X\) is a separable Banach space, (ii) \(C\) is a convex subset of \(X\) that is a Baire space such that the affine hull of \(C\) is dense in \(X,\) and (iii) \(f:C\to {\mathbb R}\) is locally Lipschitz continuous and convex. Then, for any weak\(^*\)-closed \({\mathcal M}\subset X^*,\) the following are equivalent: (1) \({\mathcal M}_f\subset {\mathcal M}.\) (2) For all \(x\in C,\) the maximization problem \[ \max_{x^*\in {\mathcal M}} [\langle x,x^*\rangle -f^*(x^*)] \] has a solution and the maximum value is equal to \(f(x).\)