Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 31 views

    Discrete Optimization: The Knapsack Problem:! Branch and Bound

    Uploaded by

    Alex Chen
    The document introduces the branch and bound algorithm for solving the knapsack problem. It discusses how branch and bound works by branching the problem into subproblems and bounding the solution to each subproblem. It also discusses how relaxing constraints, such as allowing fractional values for items in the knapsack, can provide bounds to help evaluate solutions during the search.

    Copyright:

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

    Discrete Optimization: The Knapsack Problem:! Branch and Bound

    Uploaded by

    Alex Chen
    31 views18 pages
    The document introduces the branch and bound algorithm for solving the knapsack problem. It discusses how branch and bound works by branching the problem into subproblems and bounding the solution to each subproblem. It also discusses how relaxing constraints, such as allowing fractional values for items in the knapsack, can provide bounds to help evaluate solutions during the search.

    Original Description:

    descrete optimization 3

    Original Title

    Lecture Knapsack BB

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document introduces the branch and bound algorithm for solving the knapsack problem. It discusses how branch and bound works by branching the problem into subproblems and bounding the solution to each subproblem. It also discusses how relaxing constraints, such as allowing fractional values for items in the knapsack, can provide bounds to help evaluate solutions during the search.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    31 views18 pages

    Discrete Optimization: The Knapsack Problem:! Branch and Bound

    Uploaded by

    Alex Chen
    The document introduces the branch and bound algorithm for solving the knapsack problem. It discusses how branch and bound works by branching the problem into subproblems and bounding the solution to each subproblem. It also discusses how relaxing constraints, such as allowing fractional values for items in the knapsack, can provide bounds to help evaluate solutions during the search.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 18

    Discrete

    Optimization
    The Knapsack Problem:!
    Branch and Bound

    Goals of the Lecture

    Introduce branch and bound!


    The value of relaxation

    One-Dimensional Knapsack
    maximize
    subject to

    45x1 + 48x2 + 35x3


    5x1 + 8x2 + 3x3 10
    xi 2 {0, 1} (i 2 1..3)

    Exhaustive Search
    x x x
    ? ? ?
    0

    x1 = 1

    x1 = 0

    x2 = 1

    x3 = 1
    3

    x2 = 1

    x2 = 0
    2

    x3 = 0
    3

    x3 = 1
    3

    x2 = 0

    x3 = 0
    3

    x3 = 1
    3

    x3 = 0
    3

    x3 = 1
    3

    x3 = 0
    3

    x x x

    x x x

    x x x

    x x x

    x x x

    x x x

    x x x

    x x x

    1 1 1

    1 1 0

    1 0 1

    1 0 0

    0 1 1

    0 1 0

    0 0 1

    0 0 0

    Branch and Bound

    Iterative two steps!


    branching !
    bounding!

    Branching!
    split the problem into a number of
    subproblems!
    like in exhaustive search!

    Bounding!
    find an optimistic estimate of the
    best solution to the subproblem!
    maximization: upper bound!
    minimization: lower bound
    5

    Branch and Bound

    How to find this optimistic estimate?!


    Relaxation!!
    !
    !
    !

    Optimization is the art of relaxation

    A Knapsack Model
    maximize
    subject to

    45x1 + 48x2 + 35x3


    5x1 + 8x2 + 3x3 10
    xi 2 {0, 1} (i 2 1..3)

    What can we relax?!


    we can relax the capacity constraint

    Depth-First Branch and Bound


    i

    45

    48

    35

    K = 10
    X2 = 1

    X1 = 1

    X1 = 0

    $45!
    5!
    $128

    $0!
    10!
    $83
    X2 = 0

    ---!
    -3!
    ---

    X2 = 1

    $45!
    5!
    $80
    X3=1

    $80!
    2!
    $80
    8

    Value!
    Room!
    Estimate

    $0!
    10!
    $128

    $0!
    10!
    $35

    $48!
    2!
    $83
    X3=0

    $45!
    5!
    $45

    X3=1

    ---!
    -1!
    ---

    X2 = 0

    X3=0

    $48!
    2!
    $48

    A Knapsack Model
    maximize
    subject to

    45x1 + 48x2 + 35x3


    5x1 + 8x2 + 3x3 10
    xi 2 {0, 1} (i 2 1..3)

    Can we relax something else?

    A Knapsack Model

    What if the items are bars of Belgian chocolate?!


    In that case, we could actually take a fraction
    of the bar!!
    !

    maximize
    Callebaut
    !
    subject to

    45x1 + 48x2 + 35x3

    5x1 + 8x2 + 3x3 10


    0 xi 1 (i 2 1..3)

    !
    !

    This is called the linear relaxation!


    we will come back to this later in the class!
    we relax the integrality requirement
    10

    A Knapsack Model

    Can we solve a knapsack when we can


    take parts of the items?!

    order the items by decreasing value of Vi/Wi !


    most value per kilo

    maximize
    subject to

    45x1 + 48x2 + 35x3


    5x1 + 8x2 + 3x3 10
    0 xi 1 (i 2 1..3)

    11

    A Knapsack Model

    How to solve the relaxation now?!


    select the items while the capacity is not exhausted!
    select a fraction of the last item!

    maximize
    !
    subject to
    !
    !
    !
    !

    45x1 + 48x2 + 35x3

    5x1 + 8x2 + 3x3 10


    0 xi 1 (i 2 1..3)

    In this example,!

    V1/W1 = 9, V2/W2 = 6, V3/W3 = 11.7!


    select items 3 and 1!
    select 1/4 of item 2!
    estimation: 92

    12

    A Knapsack Model

    Why is correct?
    yi
    let xi =
    vi

    maximize
    subject to

    P
    P

    y
    i21..j i
    wi
    y
    i21..j vi i

    K
    0 yi 1 (i 2 1..j)
    13

    A Knapsack Model

    Of course, if the items are archeological artifacts...

    14

    Depth-First Branch and Bound


    i

    45

    48

    35

    K = 10
    X2 = 1

    X1 = 1

    X1 = 0

    $45!
    5!
    $92

    $0!
    10!
    $77
    X2 = 0

    ---!
    -3!
    ---

    $45!
    5!
    $80
    X3=1

    $80!
    2!
    $80
    15

    Value!
    Room!
    Estimate

    $0!
    10!
    $92

    X3=0

    $45!
    5!
    $45

    Until Next Time

    16

    Citations
    Stone Foundation Tablet with Inscription of Gudea - 41221 (http://commons.wikimedia.org/wiki/File:Sumerian __Stone_Foundation_Tablet_with_Inscription_of_Gudea_-_Walters_41221_-_View_A.jpg). Artist Unknown. Walters Art Museum [Public domain, CC-BY-SA-3.0
    (http://creativecommons.org/licenses/by-sa/3.0 )], via Wikimedia Commons !

    !
    Stone Foundation Tablet with Inscription of Gudea - 41220 (http://commons.wikimedia.org/wiki/File:Sumerian_-_
    Stone_Foundation_Tablet_with_an_Inscription_of_Gudea_-_Walters_41220_-_View_A.jpg). Artist Unknown. Walters Art Museum [Public domain, CC-BY-SA-3.0
    (http://creativecommons.org/licenses/by-sa/3.0) via Wikimedia Commons !

    !
    Buddha at the Moment of Victory (http://commons.wikimedia.org/wiki/File:Thai_-_Buddha_at_the_Moment_of_Victory_-_Walters_542775.jpg ). Artist Unknown.
    Walters Art Museum [Public domain, CC-BY-SA-3.0 (http://creativecommons.org/licenses/by-sa/3.0 )], via Wikimedia Commons !

    !
    Hoxne Hoard two gold bracelets (http://commons.wikimedia.org/wiki/File:Hoxne_Hoard_two_gold_bracelets_side.JPG) by F (http://commons.wikimedia.org/
    wiki/User:F%C3%A6) [CC-BY-SA-3.0 (http://creativecommons.org/licenses/by-sa/3.0)], via Wikimedia Commons !

    !
    Ring with the engraved portrait of Ptolemy VI Philometor (http://commons.wikimedia.org/wiki/File:Ring_with_engraved_portrait_of_ Ptolemy_VI_Philometor_(3rd
    %E2%80%932nd_century_BCE)_-_2009.jpg<http://commons.wikimedia.org/wiki/File:Ring_with_engraved_
    portrait_of_Ptolemy_VI_Philometor_(3rd-2nd_century_BCE)_-_2009.jpg>) By Unknown. (Photographed by PHGCOM in 2009.) [Public domain], via Wikimedia
    Commons !

    !
    Calice du sacre Tau (http://commons.wikimedia.org/wiki/File:Calice_du_sacre_Tau.jpg) By Vassil (http://commons.wikimedia.org/wiki/User:Vassil) (Own work)
    [Public domain], via Wikimedia Commons !

    17

    Citations
    the mask of agamemnon (http://www.flickr.com/photos/rosemania/5705122218/) by Xuan Che (http://www.flickr.com/people/rosemania/) CC BY-2.0 (http://
    creativecommons.org/licenses/by/2.0/deed.en) !

    !
    Terracotta Warrior (http://www.flickr.com/photos/59627558@N00/4677378806/) by fixermark (http://www.flickr.com/photos/59627558@N00/) CC BY 2.0 (http://
    creativecommons.org/licenses/by/2.0/deed.en)!

    !
    Mmmmm Chocolate [274/366] (http://www.flickr.com/photos/sackton/8041853745) by Tim Sackton (http://www.flickr.com/photos/sackton/) CC BY-SA 2.0 (http://
    creativecommons.org/licenses/by-sa/2.0/deed.en)!

    18

    You might also like