Description
Abstract:The paper provides a statistical solution to a very general problem in non-linear programming. Specifically, it does not make any assumptions about the convexity (or, for that matter, quasiconvexity or pseudo-convexity) of any of the functions involved in the problem. On the other hand, it does not provide a guaranteed global optimum of the objective function, but is confined to: (a) The computation of a feasible and "near optimum" solution. (b) The computation of confidence limits between which the global optimum is guaranteed to lie with a specified statistical confidence coefficient.
Item Description:"January 1972."
"Research conducted through the Texas A & M Research Foundation."
"This paper is a sequel to a dissertation by Roger Pfaffenberger now at Pennsylvania State University"--Leaf 1.
Physical Description:25 leaves, 8 unnumbered leaves ; 28 cm
Bibliography:Includes bibliographical references (leaf 25).