A note on a modification of Gajda's equation (Q719662)

Let \(G\) be a locally compact abelian group, and \(m\) be a fixed Haar measure on \(G\). Denote the space of all regular bounded complex valued Borel measures on \(G\) by \(M(G)\). Further, denote the space of all \(m\)-measurable functions \(f: G\rightarrow\mathbb C\) which are bounded \(m\)-locally almost everywhere by \(L^{\infty}(G)\). Under these conventions, the main result of the paper reads as follows. Let \(\mu\in M(G)\) and let \(g\in L^{\infty}(G)\) be a Borel function which does not vanish \(m\)-locally almost everywhere and let \(f: G\rightarrow\mathbb C\) be \(m\)-measurable, not identically zero. If, for all \(x,y\in G\), functions \(f\) and \(g\) satisfy the functional equation \[ \int_G\bigl(g(x+y-s)+g(x-y+s)\bigr)d_{\mu}s=f(x)g(y), \] then \(f\) is bounded and there exist a complex constant \(c\) and continuous character \(\gamma\) such that \(g=cf\) and \(f\) is given by \[ f(x)=\int_G\bigl(\gamma(x-s)+\gamma(s-x)\bigr)d_{\mu}s. \] Conversely, if \(f\) and \(g\) are represented as above, then they fulfill the considered functional equation. The investigations are motivated by the classical d'Alembert functional equation (or, as it is also termed: the cosine equation) and a result of Gajda.