\(\mathbb Q\)-subdifferential of Jensen-convex functions (Q2497345): Difference between revisions
From MaRDI portal
Changed an Item |
Set profile property. |
||
Property / MaRDI profile type | |||
Property / MaRDI profile type: MaRDI publication profile / rank | |||
Normal rank |
Revision as of 07:23, 5 March 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | \(\mathbb Q\)-subdifferential of Jensen-convex functions |
scientific article |
Statements
\(\mathbb Q\)-subdifferential of Jensen-convex functions (English)
0 references
4 August 2006
0 references
Let \(K\) be a subfield of \(\mathbb{R}\) and \(X\) be a vector space over \(K\). A set \(D\subset K\) is called \(K\)-algebraically open if, for every \(x\in D\) and \(u\in X\), there exists \(\delta >0\) such that \(x+ru\in D\) whenever \(r\in K\cap ]-\delta ,\delta [ \). It is called \(K\)-convex if \(rx+(1-r)y\in D\) for every \(x,y\in D\) and \(r\in K\cap [ 0,1].\) A real-valued function \(f \) defined on a nonempty, \(K\)-algebraically open and \(K\)-convex set \(D\subset X\) is said to be \(K\)-convex if \(f(\lambda x+(1-\lambda )y)\leq \lambda f(x)+(1-\lambda )f(y)\) holds for every \(x,y\in D\) and \(\lambda \in [0,1]\cap K.\) Clearly, \(K\)-convex functions are Jensen-convex (i.e., they satisfy the above inequality with \(\lambda =\frac{1}{2}\));\ on the other hand, Jensen-convex functions are \(\mathbb{Q}\)-convex. The authors prove that every \(K\)-convex function is radially \(K\)-differentiable, that is, for every \(x\in D\) and \(u\in X\) the limit \(d_{K}f(x,v)\) of \(\frac{f(x+ru)-f(x)}{r}\) as \(r>0\) goes to \(0\) in \(K\) exists. The sets \(\delta _{K}f(x)\) and \(\partial _{K}f(x)\) of all additive and all \(K\)-linear, respectively, mappings \(A:X\rightarrow \mathbb{R}\) such that \(A(v)\leq d_{K}f(x,v)\) for every \(v\in X\) are called the \(K\)-subderivative and the \(K\)-subdifferential, respectively, of \(f\) at \(x\); a detailed study of these notions is presented in the paper. In particular, results analogous to those in classical convex analysis on cyclic monotonicity, maximal monotonicity and integrability of subdifferentials are given for \(K\)-subdifferentials of \(K\)-convex functions.
0 references
Jensen-convex functions
0 references
generalized subdifferential
0 references
monotonicity
0 references
cyclic monotonicity
0 references