Key points are not available for this paper at this time.
A method is described for determining an optimal straight-line segment approximation to specified functions for constrained and unconstrained endpoints.
A. Cantoni (1971) studied this question.