Convexity of functions produced by Bernstein operators of max-product kind (Q2311936)

An important property of the classical Bernstein operators, which is very useful in computer aided geometric design, is the preservation of convexity of functions. This paper is concerned with Bernstein operators of max-product type; see, e.g., \textit{B. Bede} et al. [Approximation by max-product type operators. Cham: Springer (2016; Zbl 1358.41013)]. These nonlinear operators have many interesting properties, including preservation of monotonicity and quasiconvexity of functions, but do not preserve convexity of linear functions. It is known that if a function is continuous and non-negative on \([0,1]\), then its image under a Bernstein operator of max-product type is piecewise convex. In this paper the authors show that for any continuous and non-negative function, the images generated by Bernstein operators of max-product type are not convex unless they are constant functions.