In the fifties, the theory of nonlinear programming was almost limited to problems in which the consraints and the objective functions were convex. In the sixties, it was gradually recognized that for most of the theorems ...