Documente Academic
Documente Profesional
Documente Cultură
and all , 0 1,
f [X2 + (1 )X1 ] f (X2 ) + (1 )f (X1 ) (A.1)
that is, if the segment joining the two points lies entirely above or on the graph of
f (X). Figures A.1a and A.2a illustrate a convex function in one and two dimensions,
respectively. It can be seen that a convex function is always bending upward and
hence it is apparent that the local minimum of a convex function is also a global
minimum.
Concave Function. A function f (X) is called a concave function if for any two
points X1 and X2 , and for all 0 1,
f [X2 + (1 )X1 ] f (X2 ) + (1 )f (X1 ) (A.2)
that is, if the line segment joining the two points lies entirely below or on the graph
of f (X).
Figures A.1b and A.2b give a concave function in one and two dimensions, respec-
tively. It can be seen that a concave function bends downard and hence the local
maximum will also be its global maximum. It can be seen that the negative of a con-
vex function is a concave function, and vice versa. Also note that the sum of convex
functions is a convex function and the sum of the concave functions is a concave
function. A function f (X) is strictly convex or concave if the strict inequality holds
in Eqs. (A.1) or (A.2) for any X1 = X2 . A linear function will be both convex and
concave since it satisfies both inequalities (A.1) and (A.2). A function may be con-
vex within a region and concave elsewhere. An example of such a function is shown
in Fig. A.3.
Testing for Convexity or Concavity. In addition to the definition given, the following
equivalent relations can be used to identify a convex function.
Engineering Optimization: Theory and Practice, Fourth Edition Singiresu S. Rao 779
Copyright 2009 by John Wiley & Sons, Inc.
780 Convex and Concave Functions
Figure A.1 Functions of one variable: (a) convex function in one variable; (b) concave function
in one variable.
Figure A.2 Functions of two variables: (a) convex function in two variables; (b) concave
function in two variables.
Figure A.3 Function that is convex over certain region and concave over certain other region.
Convex and Concave Functions 781
Theorem A.1 A function f (X) is convex if for any two points X1 and X2 , we have
f (X2 ) f (X1 ) + f T (X1 )(X2 X1 )
that is,
f [X1 + (X2 X1 )] f (X1 ) + [f (X2 ) f (X1 )] (A.3)
which can be seen to be the desired result. If f (X) is concave, the opposite type of
inequality holds true in (A.6).
Theorem A.2 A function f (X) is convex if the Hessian matrix H(X) = [ 2 f (X)/
xi xj ] is positive semidefinite.
It can be seen that inequality (A.6) is satisfied [and hence f (X) will be convex] if H(X)
is positive semidefinite. Further, if H(X) is positive definite, the function f (X) will be
strictly convex. It can also be proved that if f (X) is concave, the Hessian matrix is
negative semidefinite.
782 Convex and Concave Functions
The following theorem establishes a very important relation, namely, that any local
minimum is a global minimum for a convex function.
Theorem A.3 Any local minimum of a convex function f (X) is a global minimum.
Proof : Let us prove this theorem by contradiction. Suppose that there exist two different
local minima, say, X1 and X2 , for the function f (X). Let f (X2 ) < f (X1 ). Since f (X)
is convex, X1 and X2 have to satisfy the relation (A.6), that is,
f (X2 ) f (X1 ) f T (X1 )(X2 X1 ) (A.6)
or
f T (X1 )S 0 (A.9)
Example A.1 Determine whether the following functions are convex or concave.
(a) f (x) = ex
(b) f (x) = 8x 2
(c) f (x1 , x2 ) = 3x13 6x22
(d) f (x1 , x1 , x3 ) = 4x12 + 3x22 + 5x32 + 6x1 x2 + x1 x3 3x1 2x2 + 15
SOLUTION
(a) f (x) = ex : H (x) = d 2f/dx 2 = ex > 0 for all real values of x. Hence f (x) is
strictly convex.
(b) f (x) = 8x 2: H (x) = d 2f/dx 2 = 16 < 0 for all real values of x. Hence
f (x) is strictly concave.
(c) f = 2x13 6x22 :
2 f/x12 2 f/x1 x2 12x1 0
H(X) = 2 =
f/x1 x2 2 f/x22 0 12
f/x12
2
2 f/x1 x2 2 f/x1 x3
f/x1 x2 2 f/x22
2
2 f/x2 x3
H(X) =
2 f/x1 x3 2 f/x2 x3 2 f/x32
8 6 1
= 6 6 0
1 0 10
Here the principal minors are given by
|8| = 8 > 0
8 6
6 6 = 12 > 0
8 6 1
6 6 0 = 114 > 0
1 0 10
and hence the matrix H(X) is positive definite for all real values of x1 , x2 , and
x3 . Therefore, f (X) is a strictly convex function.