Scribd Logo
Upload

Welcome to Scribd!

  • 73 views

    1024 Lagrange Multipliers

    Uploaded by

    Vico Satriani
    This document discusses using Lagrange multipliers to solve optimization problems with inequality constraints. It provides an example of finding the minimum of f(x)=x^2 where 1≤x≤2. To apply Lagrange multipliers, the inequality constraints are converted to equality constraints by introducing slack variables s and t. This results in five Lagrange multiplier equations being solved to determine that the minimum occurs at x=1.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    1024 Lagrange Multipliers

    Uploaded by

    Vico Satriani
    73 views2 pages
    This document discusses using Lagrange multipliers to solve optimization problems with inequality constraints. It provides an example of finding the minimum of f(x)=x^2 where 1≤x≤2. To apply Lagrange multipliers, the inequality constraints are converted to equality constraints by introducing slack variables s and t. This results in five Lagrange multiplier equations being solved to determine that the minimum occurs at x=1.

    Original Description:

    materi persamaan lagrange untuk optimasi

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document discusses using Lagrange multipliers to solve optimization problems with inequality constraints. It provides an example of finding the minimum of f(x)=x^2 where 1≤x≤2. To apply Lagrange multipliers, the inequality constraints are converted to equality constraints by introducing slack variables s and t. This results in five Lagrange multiplier equations being solved to determine that the minimum occurs at x=1.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    73 views2 pages

    1024 Lagrange Multipliers

    Uploaded by

    Vico Satriani
    This document discusses using Lagrange multipliers to solve optimization problems with inequality constraints. It provides an example of finding the minimum of f(x)=x^2 where 1≤x≤2. To apply Lagrange multipliers, the inequality constraints are converted to equality constraints by introducing slack variables s and t. This results in five Lagrange multiplier equations being solved to determine that the minimum occurs at x=1.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 2

    MA 1024 – Lagrange Multipliers

    for Inequality Constraints


    Here are some suggestions and additional details for using Lagrange mul-
    tipliers for problems with inequality constraints.
    Statements of Lagrange multiplier formulations with multiple equality
    constraints appear on p. 978-979, of Edwards and Penney’s Calculus Early
    Transcendentals, 7th ed. Refer to them. Note the condition that the gradients
    of the constraints (e.g., ∇g and ∇h in Theorem 2, p. 978) must be nonzero
    and nonparallel.
    Here’s an example with inequality constraints: find the minimum of
    f (x) = x2 for 1 ≤ x ≤ 21 .
    Written separately, the inequality constraints are x−1 ≥ 0 and 2−x ≥ 0.
    To convert them to equality constraints, introduce two new variables s
    and t and corresponding equality constraints:

    g1 (x, s, t) = x − 1 − s2 = 0
    g2 (x, s, t) = 2 − x − t2 = 0

    Squaring the new variables insures that these terms are non-negative, thereby
    capturing the inequality constraints. The variables s and t are called slack
    variables because they take up the slack in the inequalities.
    The revised problem is: minimize f (x) = x2 subject to g1 (x, s, t) =
    x − 1 − s2 = 0 and g2 (x, s, t) = 2 − x − t2 = 0.
    The Lagrange multiplier formulation is: solve

    g1 (x, s, t) = x − 1 − s2 = 0
    g2 (x, s, t) = 2 − x − t2 = 0
    ∇f (x) = λ1 ∇g1 (x, s, t) + λ2 ∇g2 (x, s, t)

    The gradient operates on the three variables (x, s, t); i.e., ∇ = h∂x , ∂s , ∂t i.
    Hence,

    ∇f = h2x, 0, 0i
    ∇g1 = h1, −2s, 0i
    ∇g2 = h−1, 0, −2ti
    1
    Of course, the minimum occurs at x = 1.

    1
    Hence, the five Lagrange multiplier equations are

    x − 1 − s2 = 0 (1)
    2 − x − t2 = 0 (2)
    2x = λ1 − λ2 (3)
    0 = −2sλ1 (4)
    0 = −2tλ2 (5)

    There are two possibilities with each inequality constraint, active – up


    against its limit – or inactive, a strict inequality. If the constraint is active,
    the corresponding slack variable is zero; e.g., if x − 1 = 0, then s = 0. The
    inequality constraint is actually functioning like an equality, and its Lagrange
    multiplier is nonzero. If the inequality constraint is inactive, it really doesn’t
    matter; its Lagrange multiplier is zero.
    Equations (4) and (5) offer these alternatives:

    From (4): either s 6= 0 and λ1 = 0, in which case x = 1 + s2 > 1, or s = 0


    and x = 1, from (1).

    From (5): either t 6= 0 and λ2 = 0, in which case x = 2 − t2 < 2, or t = 0


    and x = 2, from (2).

    Hence, according to equations (1), (2), (4) and (5), just one of the following
    is possible:

    Case I: x = 1.

    Case II: x = 2.

    Case III: 1 < x < 2 and λ1 = λ2 = 0.

    In Case III, equation (3) forces x = (λ1 − λ2 )/2 = 0, violating 1 < x < 2. So
    III is impossible.
    Which of I and II gives the lower value to f ? Of course, it’s Case I; x = 1
    gives f (x) = x2 its minimum value on 1 ≤ x ≤ 2.
    Checking points to eliminate impossibilities and to locate the actual ex-
    tremum is typical of Lagrange multipliers.

    You might also like