Scribd Logo
Upload

Welcome to Scribd!

  • 306 views

    Convex and Concave Functions

    Uploaded by

    rnmukerjee
    describes basic symptomps of convex and concave functions and explains how to determine the nature of a function.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    Convex and Concave Functions

    Uploaded by

    rnmukerjee
    306 views3 pages
    describes basic symptomps of convex and concave functions and explains how to determine the nature of a function.

    Original Title

    Convex and concave functions

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    describes basic symptomps of convex and concave functions and explains how to determine the nature of a function.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    306 views3 pages

    Convex and Concave Functions

    Uploaded by

    rnmukerjee
    describes basic symptomps of convex and concave functions and explains how to determine the nature of a function.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 3

    APPENDIX

    CONVEX AND CONCAVE FUNCTIONS


    Convex Function

    A function of n variables (x) dened on a convex set D is said to be a convex function if and only if for any two points x (1) and x (2) D and 0 1, [ x (1) (1 ) x (2)] (x (1)) (1 )(x (2)) Figure B.1 illustrates the denition of a convex function of a single variable.
    Properties of Convex Functions

    1. The chord joining any two points on the curve always falls entirely on or above the curve between those two points. 2. The slope or rst derivative of (x) is increasing or at least nondecreasing as x increases. 3. The second derivative of (x) is always nonnegative for all x in the interval. 4. The linear approximation of (x) at any point in the interval always underestimates the true function value. 5. For a convex function, a local minimum is always a global minimum. Figure B.2 illustrates property 4. The linear approximation of at the point (x; x 0), is obtained by ignoring the second and other higher x 0, denoted by order terms in the Taylor series expansion
    x; x 0) (x 0) (x 0)(x x 0) (

    For a convex function, property 4 implies that

    648

    Engineering Optimization: Methods and Applications, Second Edition. A. Ravindran, K. M. Ragsdell and G. V. Reklaitis 2006 John Wiley & Sons, Inc. ISBN: 978-0-471-55814-9

    CONVEX AND CONCAVE FUNCTIONS

    649

    Figure B.1. Convex function.

    (x) (x 0) (x 0)(x x 0)

    for all x

    The gradient of a function (x1, . . . , xn) is given by


    (x1, . . . , xn)

    , ,..., x1 x2 xn

    The Hessian matrix of a function (x1, . . . , xn) is an n n symmetric matrix given by H(x1, . . . , xn)

    2 2 xi xj

    Figure B.2. Linear approximation of a convex function.

    650

    CONVEX AND CONCAVE FUNCTIONS

    Test for Convexity of a Function. A function is convex function if the

    Hessian matrix of is positive denite or positive semidenite for all values of x1, . . . , xn.
    Concave Function

    A function (x) is a concave function in D if and only if (x) is a convex function in D.


    Test for Concavity of a Function. A function is concave if the Hessian

    matrix of is negative denite or negative semidenite for all values of x1, . . . , xn.
    Example B.1
    2 2 2 (x1, x2, x3) 3x 1 2x 2 x3 2 x1x2 2 x1x3

    2 x2 x3 6x1 4 x2 2 x3 f(x1, x2, x3)

    6 2 4 H(x1, x2, x3) 2 2 2

    6 x1 2 x2 2 x3 6 4 x2 2 x1 2 x3 4 2 x3 2 x1 2 x2 2
    2

    2 2

    To show that is a convex function, we test for whether H is positive denite or positive semidenite. Note that: 1. H is symmetric. 2. All diagonal elements are positive. 3. The leading principal determinants are
    6 0

    6 2 20 0 4

    H 16 0

    Hence H is a positive-denite matrix, which implies is a convex function. (As a matter of fact, when H is positive denite, is said to be strictly convex with a unique minimum point.)

    You might also like