Some remarks stemming from Ulam's problem about nearly additive mappings (Q1291168)

The author considers the following problems: Let \((S,+)\) be a semigroup and let \(\rho:S \to \mathbb R\) be a given control function. Assume that \(F:S \to \mathbb R\) is a ''nearly additive'' mapping in either of the following ways: \[ | F(x+y)-F(x)-F(y)| \leq \rho(x)+\rho(y)-\rho(x+y), \quad x,y \in S,\tag{1} \] \[ \left| \sum_{i=1}^n F(x_i)-\sum_{j=1}^m F(y_j) \right| \leq \sum_{i=1}^n \rho(x_i)+\sum_{j=1}^m \rho(y_j),\tag{2} \] for all \(n\) and \(m\) whenever \(x_i\) and \(y_j\) are elements of \(S\) such that \(\sum_{i=1}^n x_i=\sum_{j=1}^m y_j\). Must \(F\) be ``near'' to an additive mapping \(A:S \to \mathbb R\) in the sense that \[ | F(x)-A(x)| \leq K\rho(x)\text{ for some }K \text{ and all }x \in S ?\tag{3} \] The main result of the paper is the following. Theorem. Let \((S,+)\) be a semigroup and let \(K\) be a fixed number. The following statements are equivalent: (a) For every \(\rho\) and every \(F\) satisfying (1) there is an additive \(A\) fulfilling (3). (b) For every \(\rho\) and every \(F\) satisfying (2) there is an additive \(A\) fulfilling (3). Moreover, for \(K=1\) (a) and (b) are equivalent to (c) For every \(\alpha, \beta:S \to \mathbb R\) such that \(\alpha\) is superadditive, \(\beta\) is subadditive and \(\alpha \leq \beta\), there exists an additive \(A\) separating \(\alpha\) from \(\beta\), i.e., \(\alpha(x)\leq A(x) \leq \beta(x)\).