Polynomial approximations of continuous functions with unbounded sign-sensitive weight (Q2462057)

The author defines the \(p\)-norm of a function \(f\) on the closed interval \(\Delta= [a,b]\) as follows \[ |f|_p= |f|_{p,\Delta}= \sup\{|(f, p)(x)|: x\in\Delta, \] where \(|(f, p)(x)|= f^+(x) p_+(x)+ f^-(x) p_2(x)\), \(p_+(x)\) and \(p_-(x)\) being components of the sign-sensitive weight \(p= (p_-, p_+)\). The \(p\)-norm generalizes the ordinary uniform norm and is sublinear functional on the set of functions bounded on \(\Delta\). The author considers the problem of approximation theory in the \(p\)-norm. These problems differ from the problems of approximation theory in the ordinary uniform norm ``just in the case of arbitrary weights where the components of the weights on the given closed intervals can be discontinuous, unbounded, or can vanish on its arbitrary subsets'' (quoted from the author's introduction). In a previous paper [Anal. Math. 28, No. 3, 205--230 (2002; Zbl 1017.41007)] the author has given the necessary and sufficient conditions of polynomial approximation of a given finite function \(f\) in the \(p\)-norm on the closed interval \(\Delta\) in the case when the set \(F= \Delta^\infty_-\cap \Delta^\infty_+\) is empty, where \(\Delta^\infty_-\) and \(\Delta^\infty_+\) are the sets of those points from \(\Delta\) in any neighborhood of which \(p_-(x)\) and \(p_+(x)\) are unbounded. In the present paper the problem of the polynomial approximation of continuous functions is considered for the case where the set \(F\) is finite. Finally, an analog of the module of continuity in the \(p\)-norm is defined through the divided differences.