Appl Math Optim 20:19-31 (1989)
Applied Mathematics
and Optimization
© 1989Springer-VerlagNew York Inc.
Sufficiency Conditions with Minimal Regularity Assumptions*
Vera Zeidan
Department of Applied Mathematics, University of Waterloo, Waterloo,
Ontario, Canada N2L 3G1
Communicated by W. Fleming
Abstract. The aim of this article is to provide second-order sufficiency
criteria that extend known ones in [6] and [7] to the case where the control
candidate and some of the data are merely essentially bounded, and/or the
control set U is any convex subset of R m. In the classical setting, where a
restriction on the velocity is imposed (~(t)~ U), it is shown that when U is
compact the known strengthened Weierstrass condition is equivalent to the
Weierstrass condition with strict inequality.
I. Introduction
1.1. The Problem
Given an interval [a, b], a point A in R ~, a subset U of R m, functions f and g
mapping [a, b] x R" x R "~ to R n and R, respectively, and a function ~¢ from R n to
R, the optimal control problem is defined to be:
+ b
(C) Minimize J(X, U) "= ~(x(b)) ~a g(t, x(t), u(t)) dt over all pairs (x, u)
of absolutely continuous functions x: [a, b] ~ R n and measurable func-
tions u: [a, bier m such that
Yc(t) =f(t,x(t), u(t)) a.e.,
x(a)=A,u(t)~U a.e.;
such a pair is called feasible.
* The research was supported by an NSERC Grant and by GNAFA of CNR, which are gratefully
acknowledged.