On regular solutions of functional equations (Q1074039)

This is one in a series of papers by the author studying regular solutions of the following rather general class of functional equations [Publ. Math. 26, 17-35 (1979; Zbl 0433.39012) and Aequationes Math. 25, 52-66 (1982; Zbl 0517.39009)]: \[ (1)\quad f(x)=h[x,y,f_ 0(y),f_ 1(g_ 1(x,y)),...,f_ n(g_ n(x,y))]. \] Here \(f,f_ 0,f_ 1,...,f_ n\) are unknown functions and \(h,g_ 1,...,g_ n\) are given functions, where the domains and ranges of these functions are topological, measure and metric spaces. The author proves - for appropriate functions \(h,g_ 1,...,g_ n\)- results of the following type: 1. Measurable solutions \(f_ 1,...,f_ n\) of (1) imply the continuity of f. 2. Almost open solutions \(f_ 1,...,f_ n\) of (1) imply the continuity of f. 3. Continuous solutions \(f_ 1,...,f_ n\) of (1) imply that f is locally Lipschitz (This statement is proven only in two special cases of (1) and thus statement 3. is an open problem). 4. Locally Lipschitz solutions \(f_ 1,...,f_ n\) of (1) imply the continuously differentiability of f. 5. p times continuously differentiable solutions \(f_ 1,...,f_ n\) of (1) imply that f is \(p+1\) times continuously differentiable. It is remarkable that in all these results no assumptions about \(f_ 0\) are needed. The importance of these results is that the functional equation (1) can be solved in the following way (say, in the special case \(f=f_ 1=...=f_ n)\) (a) Every solution of (1) satisfying a weak regularity property is infinitely many times differentiable. (b) Reduce the functional equation to a differential equation and solve it (of course, (b) depends on the special form of the functional equation). Finally, the justification, for considering the general functional equation (1), is given by several illustrating applications.