A minimax formula for the best natural \(C([0,1])\)-approximate by nondecreasing functions (Q2371932)

Let \(a,b\in [0,1]\), \(a<b\), \(f\in C([0,1])\) and \(m_\infty (a,b)=\frac{1}{2}(\max _{[a,b]}f+ \min_{[a,b]}f).\) The functions \(f_\infty (x):=\sup_{a<x}\inf_{b>x} m_\infty (a,b)\) and \(f^\infty (x):=\inf_{b>x}\sup_{a<x}m_\infty (a,b)\), \(x\in (a,b),\) are introduced. By \textit{R. Darst} and \textit{S. Sahab} [J. Approximation Theory 38, 9--27 (1983)] was proved that for \(f\in C({0,1})\) the best \(L_p([0,1])\) approximants \((1<p<\infty )\) to \(f\) by nondecreasing functions converge uniformly as \(p\rightarrow \infty \) to \(f_*\), which is a best \(C([0,1])\)-approximant to \(f\) by nondecreasing functions. In the paper it is proved that \(f_*=f_\infty =f^\infty \). Besides some additional minimization properties of \(f_\infty \) are established.