A Kuczma-type functional inequality for error and complementary error functions (Q2350288)

The error, resp. complementary error functions are defined by \[ \text{erf}(x)= {2\over\sqrt{\pi}} \int^x_0 e^{-t^2}\,dt, \] and \(\text{erf}_c(x)= 1-\text{erf}(x)\). Inspired by an interesting functional equation by \textit{M. Kuczma} [Rocz. Nauk.-Dydakt., Pr. Mat. 13, 197--213 (1993; Zbl 0964.39501)], namely \(x+ F(y+G(x))= y+F(x+ G(y))\), the author proves that for all real numbers \(0\leq x<y\), one has the inequality \(x+\text{erf}(y+ \text{erf}_c(y))\). The proof is based on computations with Maple, and many particular subintervals. The referee feels, this is not a proof from the ``book''.