On a variant of Wilson's functional equation on groups (Q704877)

Let \(G\) be a (nonabelian) group. The functional equation \[ f(xy)+f(y^{-1}x)=2f(x)g(y),\quad x,y\in G,\tag{1} \] with unknown functions \(f,g:G\to\mathbb{C}\), is studied. Equation (1) is a variant of the Wilson's one \[ f(xy)+f(xy^{-1})=2f(x)g(y),\quad x,y\in G.\tag{2} \] On some groups the classes of solutions of (1) and (2) may be different. It is the main aim of the paper to describe the relations between those classes and to find some conditions ensuring that a solution of (1) satisfies (2) as well. In a particular case \(g=1\), (1) implies (2) (which becomes the Jensen's equation), as it was shown by \textit{C. T. Ng} [Aequationes Math. 62, 143--159 (2001; Zbl 1004.39018)]. In a particular case \(f=g\), (1) is equivalent to (2) (which is the d'Alembert's equation in this case) on any group.