Almost convexity on Abelian groups (Q623414): Difference between revisions

Almost convexity is defined in the sense of abstract \(\sigma\)-ideals. A nonempty family \(\mathcal{I}\) of subsets of a set \(X\) is called a \(\sigma\)-ideal if it is hereditary: \[ A\in \mathcal{I}\text{ and }B\subset A\text{ imply }B\in \mathcal{I} \] and \(\sigma\)-additive: \[ A_{n}\in \mathcal{I}\text{ for }n\in \mathbb{N}\text{ implies }\bigcup _{n=1}^{\infty }A_{n}\in \mathcal{I}. \] A \(\sigma\)-ideal is proper if \(X\notin \mathcal{I}\). A condition \( \mathcal{P}\) is satisfied for \(\mathcal{I}\)-a.a. elements of \(X\) if there exists a set \(A\in \mathcal{I}\) such \(\mathcal{P}\) is satisfied for every \( x\in X\backslash A.\) Let \(A\) be a subset of an Abelian group \(G\) and \(\mathcal{I}\) be a proper \( \sigma\)-ideal in \(G\times G\). A function \(f:A\rightarrow \mathbb{R}\) is said to be convex if \[ 2f(x)\leq f(x+h)+f(x-h) \tag{1} \] for all \(x,h\in G\) satisfying \(x,x+h,x-h\in A.\) The function \(f\) is said to be \(\mathcal{I}\)-almost convex if inequality \((1)\) holds for \(\mathcal{I}\)-a.a. \((x,h)\in G\times G\) with \(x,x+h,x-h\in A.\) The author generalizes some known results of the following type: for a given \(\mathcal{I}_{2}\)-almost convex function \(f\) on \(A\) there exists a convex function \(g\) on \(A\) which equals \(f\) for \(\mathcal{I}_{1}\)-a.a. \(x\in \) \(A.\) By their nature, all the results are rather complicated. The hypotheses have to include conditions on \(G,A,\mathcal{I}_{1}\) and \(\mathcal{I}_{2}.\)