Existence, construction and extension of continuous solutions of an iterative equation with multiplication (Q6051322)

The main goal of the paper is to deal with the equation \[ (g(x))^{{\alpha}_1}(g^2(x))^{{\alpha}_2} \ldots (g^n(x))^{{\alpha}_n} = G(x) \] on the real line (or on intervals), where \(G\) is given and \(g\) is unknown. The power \(g^n\) denotes the \(n\)-th iterate: \(g^0 = {\mathrm{id}}\) and inductively \(g^{n+1} = g^n \circ g\) for \(n \in {\mathbb{N}}\). In the case that domain and codomain are the positive reals \({\mathbb{R}}_{+} = {]0,{\infty}[}\), the equation (by use of logarithm) can be transformed to \[ {\alpha}_1\,f(x) + {\alpha}_2\,f^2(x) + \ldots + {\alpha}_n\,f^n(x) = F(x). \] For this equation several results available in the literature are adapted rigorously in order to give answers on existence, uniqueness (or non-uniqueness) of solutions, especially continuous or Lipschitzian solutions, on \({\mathbb{R}}_{+}\). However, it turns out that an extension to continuous solutions on \({\mathbb{R}}\) causes problems. In many cases it is impossible, but there are cases where such a solution exists, and in this case we have a solution of the original equation with multiplication. The paper ends with some examples, where suitable construction methods provide solutions. I enjoyed the clear and transparent writing of the paper with rigorous, but easily readable proofs.