+ All Categories
Home > Documents > Chapter 4 ITIS

Chapter 4 ITIS

Date post: 14-Apr-2018
Category:
Upload: mannybanwait
View: 231 times
Download: 0 times
Share this document with a friend

of 40

Transcript
  • 7/27/2019 Chapter 4 ITIS

    1/40

    Chapter 4 -

    Optimization andMathematical

    Programming

    2002 South-Western/Thomson Learning

  • 7/27/2019 Chapter 4 ITIS

    2/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    Startron Data

    Computer Model

    Desktop Server Notebook

    Prof i t per unit $75 $145 $125

    Assembly

    t ime (hrs ) 0.5 0.75 1.5 12

    Software

    ins tal lat ion (hrs ) 0.25 0.4 0.3 3

    Test ing (hrs ) 1 1.5 1 20

    Packaging (hrs) 0.1 0.1 0.2 2

    Number of

    Workers

    ANDeach type should be at least 20% of total production

  • 7/27/2019 Chapter 4 ITIS

    3/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    Startrons Production

    Planning Problem

    Object ive:Maxim ize prof i t =75DT +145SV +125NB

    Subject to:

    0.5DT +0.75SV +1.5NB 96 (avai lable assembly hou rs)0.25DT +0.4SV +0.3NB 24 (avai lable so ftware instal lat ion hou rs)1DT +1.5SV +1NB 160 (avai lable test ing hou rs)0.1DT +0.15SV +0.2NB 16 (avai lable packaging hou rs)0.8DT- 0.2SV- 0.2NB 0 (each typ e sho uld accoun t for at

    least 20% of the total produ ct ion )

    DT 0; SV 0; NB 0 (Non negat ivi ty cons traint)

  • 7/27/2019 Chapter 4 ITIS

    4/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    The Modeling Process

    for Optimization Studies

    Optimizat ion:the name of a fami ly of tools

    designed to help solve prob lems invo lv ing

    l im i ted resources d ist r ibuted among var iousact iv i t ies to opt im ize a measu rable goal

    A l im i ted quant i ty of resou rces is avai lable

    The resources are used in the product ion o f

    produc ts or serv ices

    There are two or m ore possib le solut ions

    Each act ivi ty y ields a return related to the goal

    The al location o f resources is cons trained

  • 7/27/2019 Chapter 4 ITIS

    5/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    Linear Programming

    Output-m ix prob lems:how much o f wh ich

    ou tputs (product or serv ice m ix) should be

    planned to best ob tain the ob ject ive (i .e.,maximum prof i t ) so (con strained) resou rces

    can be approp riately al loc ated

    Blending prob lems:what is the best b lend

    of available (cons trained) components that

    w i l l ach ieve the ob ject ive (i .e., low cost)

    whi le meet ing speci f icat ion

  • 7/27/2019 Chapter 4 ITIS

    6/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    Formulating the Linear

    Programming Model

    Al l l inear programm ing(LP) models havethese s ix components :

    Decis ion Variables

    Object ive Func t ion

    Prof i t or Cost Coeff ic ients

    Constraints Cons traint Coeff ic ients

    Righ t-Hand -Side Constants

  • 7/27/2019 Chapter 4 ITIS

    7/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    The Blending Problem:

    Minimization Problem

    A paint can be prod uced us ing two ingredients

    Alpha and Beta

    Both ingredients contribute towards the paintsquality, determined by its granularityand density

    One canno t com prom ise on the paint qual ity as

    determ ined by the minim um levels of densi ty and

    granular i ty

    You w i l l need to m inimize the cost of ing redients

    whi le at the same time achieving th e desired quali ty

  • 7/27/2019 Chapter 4 ITIS

    8/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    The Blending Problem

    Decisio n Variablesx1 = Quanti ty of Alpha to be inclu ded, in oun ces,

    x2 = Quanti ty of Beta to be includ ed, in ounces

    Constra ints:

    Each d rum mus t have a granular i ty rat ing of 300

    One ounce of x1inc reases the granulari ty by 1 unitOne ounce of x2increases the granular i ty by 1 uni t

    Each d rum must h ave a dens ity rat ing o f 250 un itsOne ounce of x1increases the densi ty b y 3 un i ts

    Oneounce of x2incr eases the density b y 0 units (no effect on

    densi ty)

    Costs

    x1 = $45.00 per ounc ex2 = $12.00 per ounc e

  • 7/27/2019 Chapter 4 ITIS

    9/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    The Blending Problem: Diskote

    Minim ize z =45x1 +12x2

    Subject to:

    1x1 + 1x2 300 (granu lar i ty cons traint)3x1 + 0x2 250 (dens i ty cons traint)x1, x2 0 (nonnegat iv i ty cons traint)

  • 7/27/2019 Chapter 4 ITIS

    10/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97FALL 2010

    Assumptions of LP

    Certainty

    Linear Object ive Funct ion

    Linear Cons traints : Addi t iv i ty

    Independence

    Proport ional i ty

    Nonnegat iv i ty

    Divis ib i l i ty

  • 7/27/2019 Chapter 4 ITIS

    11/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2 40 (labor con straint)

    1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

  • 7/27/2019 Chapter 4 ITIS

    12/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2 40 (labor con straint)

    1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    x1

    Product ion of leather chairs

    | | | |10 20 30 40 45

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs 40

    30

    20

    10

    0

    15

    A

    B

    C

    DE

    F

    G

    O

    Feasible A rea

  • 7/27/2019 Chapter 4 ITIS

    13/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z = 300x1 + 250x2

    Subject to:

    2x1

    + 1x2 40 (labor constraint)

    1x1 + 3x2 45 (machine t ime cons traint)1x1 + 0x2 12 (market ing c ons traint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    x1

    Product ion of leather chairs

    | | | |10 20 30 40 45

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs 40

    30

    20

    10

    0

    15

    A

    B

    C

    DE

    F

    G

    O

    Feasible A rea

    Solut ion Total Prof i t

    Poin t (Corn er) Coo rdin ates 300x1 +250x2 = z

    O (0, 0) 300 (0) + 250 (0) = 0

    C (0, 15) 300 (0) + 250 (15) = $3,750

    E (12, 0) 300 (12) + 250 (0) = $3,600

    G (12, 11) 300 (12) + 250 (11) = $6,350

  • 7/27/2019 Chapter 4 ITIS

    14/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z = 300x1 + 250x2

    Subject to:

    2x1

    + 1x2 40 (labor constraint)

    1x1 + 3x2 45 (machine t ime cons traint)1x1 + 0x2 12 (market ing c ons traint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    x1

    Product ion of leather chairs

    | | | |10 20 30 40 45

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs 40

    30

    20

    10

    0

    15

    A

    B

    C

    DE

    F

    G

    O

    Feasible A rea

    Solut ion Total Prof i t

    Poin t (Corn er) Coo rdin ates 300x1 + 250x2 = z

    O (0, 0) 300 (0) + 250 (0) = 0

    C (0, 15) 300 (0) + 250 (15) = $3,750

    E (12, 0) 300 (12) + 250 (0) = $3,600

    G (12, 11) 300 (12) + 250 (11) = $6,350

    Maximum

  • 7/27/2019 Chapter 4 ITIS

    15/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2 40 (labor con straint)

    1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

  • 7/27/2019 Chapter 4 ITIS

    16/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2 40 (labor con straint)

    1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    x1

    Product ion of leather chairs

    | | |5 10 15

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs

    15

    10

    5

    0

    12

    C

    E

    G

    O

  • 7/27/2019 Chapter 4 ITIS

    17/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2 40 (labor con straint)

    1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    x1

    Product ion of leather chairs

    | | |5 10 15

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs

    15

    10

    5

    0

    12

    C

    E

    G

    O

    M

    N

    z = 300x1 + 250x2 = 1,500

  • 7/27/2019 Chapter 4 ITIS

    18/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    The Graphical Method

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2 40 (labor con straint)

    1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    x1

    Product ion of leather chairs

    | | |5 10 15

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs

    15

    10

    5

    0

    12

    C

    E

    G

    O

    M

    N

    z = 300x1 + 250x2 = 1,500

    G

    V

    W

    z = 6,350

  • 7/27/2019 Chapter 4 ITIS

    19/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Slack and Surplus Variables

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1 + 1x2 + s1 =40 (labor con straint)

    1x1 + 3x2 + s2 =45 (machine t ime con straint)1x1 + 0x2 + s3 =12 (market ing constraint)

    x1, x2 0 (non negat ivi ty con straint)

    The Outpu t-Mix Prob lem

  • 7/27/2019 Chapter 4 ITIS

    20/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Slack and Surplus Variables

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1 + 1x2 + s1 =40 (labor con straint)

    1x1 + 3x2 + s2 =45 (machine t ime con straint)1x1 + 0x2 + s3 =12 (market ing constraint)

    x1, x2 0 (non negat ivi ty con straint)

    The Outpu t-Mix Prob lem

    Slack variables have been added to each cons traint

    At the opt imal solu t ion o f x1 =12, x2 =11, the slackvariables have the fol low ing values:

    s1 =5 s2 =0 s3 =0

  • 7/27/2019 Chapter 4 ITIS

    21/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Slack and Surplus Variables

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1 + 1x2 + s1 =40 (labor con straint)

    1x1 + 3x2 + s2 =45 (machine t ime con straint)1x1 + 0x2 + s3 =12 (market ing constraint)

    x1, x2 0 (non negat ivi ty con straint)

    The Outpu t-Mix Prob lem

    Slack variables have been added to each cons traint

    At the opt imal solu t ion o f x1 =12, x2 =11, the slackvariables have the fol low ing values:

    s1 =5 s2 =0 s3 =0

    There are 5 unused hours

    of labo r capaci ty

    The avai lable machinet ime is ful ly u t i l ized

    The market ing constra int

    is exact ly m et

    Analysis

  • 7/27/2019 Chapter 4 ITIS

    22/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Sensitivity Analysis:

    Objective Function

    When does a pr ice (pro f i t )

    decrease of an output jus t i fyd iscont inu ing or reduc ing the

    quant i ty?

    How much o f a pr ice (pro f i t)

    inc rease jus t i f ies increasing the

    quant i ty in the op t imal so lut ion?

  • 7/27/2019 Chapter 4 ITIS

    23/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Changes in the Coefficients

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2

    40 (labor con straint)1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

  • 7/27/2019 Chapter 4 ITIS

    24/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2

    40 (labor con straint)1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    Changes in the Coefficients

    x1

    Product ion of leather chairs

    | | | |20 45

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs 40

    0

    15

  • 7/27/2019 Chapter 4 ITIS

    25/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2

    40 (labor con straint)1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    Changes in the Coefficients

    x1

    Product ion of leather chairs

    | | | |20 45

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs 40

    0

    15

    A

    B

    C

    DE

    F

    G

    O

  • 7/27/2019 Chapter 4 ITIS

    26/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2

    40 (labor con straint)1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    Changes in the Coefficients

    x1

    Product ion of leather chairs

    | | | |20 45

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs 40

    0

    15

    A

    B

    C

    DE

    F

    G

    O

    K

    L

  • 7/27/2019 Chapter 4 ITIS

    27/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2

    40 (labor con straint)1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    Changes in the Coefficients

    x1

    Product ion of leather chairs

    | | | |20 45

    x2

    Pro

    duc

    tion

    offabr

    ic

    cha

    irs 40

    0

    15

    A

    B

    C

    DE

    F

    G

    O

    K

    L

    M

    N

  • 7/27/2019 Chapter 4 ITIS

    28/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2

    40 (labor con straint)1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    Changes in the Coefficients

    x1

    Product ion of leather chairs

    | | | |20 45

    x2

    Pro

    duc

    tion

    offabric

    cha

    irs 40

    0

    15

    A

    B

    C

    DE

    F

    G

    O

    K

    L

    M

    N

    Changes in the coeff ic ients change

    the slope of the ob ject ive func t ion

    This w i l l change the value of the

    solut io n, bu t not n ecessar ily the

    opt imal point

  • 7/27/2019 Chapter 4 ITIS

    29/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =300x1 +250x2

    Subject to:

    2x1

    + 1x2

    40 (labor con straint)1x1 + 3x2 45 (machine t ime con straint)1x1 + 0x2 12 (market ing constraint)x1, x2 0 (non negat ivi ty constraint)

    The Outpu t-Mix Problem

    Changes in the Coefficients

    x1

    Product ion of leather chairs

    | | | |20 45

    x2

    Pro

    duc

    tion

    offabric

    cha

    irs 40

    0

    15

    A

    B

    C

    DE

    F

    G

    O

    K

    L

    M

    N

    Changes in the coeff ic ients change

    the slope of the ob ject ive func t ion

    This w i l l change the value of the

    solut io n, bu t not n ecessar ily the

    opt imal point

    Upper and Lower Lim i ts

    There are l im its to the extent

    coeff ic ients can be changed

    before the opt imal so lut ion

    poin t changes

  • 7/27/2019 Chapter 4 ITIS

    30/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Changes in Right-Hand Side Values

    Maxim ize z =3x1 +4x2

    Subject to:

    3x1 + 5x2 152x1 + 1x2 80x1 + 1x2 2x1, x2 0

  • 7/27/2019 Chapter 4 ITIS

    31/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Changes in Right-Hand Side Values

    x1| |4 5

    x2

    9

    8

    2

    0

    Exhibit 4.33Maxim ize z =3x1 +4x2

    Subject to:

    3x1 + 5x2 152x1 + 1x2 80x1 + 1x2 2x1, x2 0

  • 7/27/2019 Chapter 4 ITIS

    32/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Changes in Right-Hand Side Values

    x1| |4 5

    x2

    9

    8

    2

    0

    Exhibit 4.33

    1

    2

    3

    A B

    C

    D

    F

    Isoprof i t

    l ine

    Maxim ize z =3x1 +4x2

    Subject to:

    3x1 + 5x2 152x1 + 1x2 80x1 + 1x2 2x1, x2 0

  • 7/27/2019 Chapter 4 ITIS

    33/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Changes in Right-Hand Side Values

    x1| |4 5

    x2

    9

    8

    2

    0

    Exhibit 4.33

    1

    2

    3

    A B

    C

    D

    F

    Isoprof i t

    l ine G

    E

    S

    T

    Maxim ize z =3x1 +4x2

    Subject to:

    3x1 + 5x2 152x1 + 1x2 80x1 + 1x2 2x1, x2 0

  • 7/27/2019 Chapter 4 ITIS

    34/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =3x1 +4x2

    Subject to:

    3x1 + 5x2 152x1 + 1x2 80x1 + 1x2 2x1, x2 0

    Changes in Right-Hand Side Values

    x1| |4 5

    x2

    9

    8

    2

    0

    Exhibit 4.33

    1

    2

    3

    A B

    C

    D

    F

    Isoprof i t

    l ine G

    E

    S

    T

    Changin g the RHS value shif ts

    Cons traint 2 fur ther from the

    or ig in .

    This resul ts in a new opt imal

    so lut ion at G

  • 7/27/2019 Chapter 4 ITIS

    35/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Maxim ize z =3x1 +4x2

    Subject to:

    3x1 + 5x2 152x1 + 1x2 80x1 + 1x2 2x1, x2 0

    Changes in Right-Hand Side Values

    x1| |4 5

    x2

    9

    8

    2

    0

    Exhibit 4.33

    1

    2

    3

    A B

    C

    D

    F

    Isoprof i t

    l ine G

    E

    S

    T

    Changin g the RHS value shif ts

    Cons traint 2 fur ther from the

    or ig in .

    This resul ts in a new opt imal

    so lut ion at G

    Upper and Lower Lim i ts

    There are l im its to the extent

    RHS values can be changed

    before other cons traints

    become l imi t ing factors

  • 7/27/2019 Chapter 4 ITIS

    36/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Integer Programming

    A form of programm ing where some

    or al l of the decis ion var iables mus t

    be in tegers

    It al low s us to solve prob lems

    where indiv is ib i li ty is p resent

    Prov ides solut ion opt ions toprob lems not o therwise solvable

  • 7/27/2019 Chapter 4 ITIS

    37/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Integer Programming

    Dif fers from no rmal LP in several ways :

    Ind iv is ibi l i ty resul ts in add it ional

    constra ints

    Solut ion is general ly infer ior to theop t imal LP so lut ion

    Complex so lut ion techn iques Gomory Cu tt ing Plane Method

    Complete enumeration

    Branch and bound

    Typ ical ly a f in i te number of so lut ions

  • 7/27/2019 Chapter 4 ITIS

    38/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Integer Programming

    Integer Prog ramm ing Models:

    A ll-In teger Model

    All decis ion var iables mus t be integers

    Mixed In teger Model Only som e of the decis ion var iables are

    integers Zero -One In teger Model

    Useful for modeling yes/no decisions

  • 7/27/2019 Chapter 4 ITIS

    39/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Southern General Hospital

    Decis ion variables:

    x1 = Number of beds to add to the materni ty ward

    x2 = Number of beds to add to the cardiac ward

    Constra ints:

    Only 10 beds to tal can be added

    $210,000 to tal budget At least two of the beds must be in materni ty

  • 7/27/2019 Chapter 4 ITIS

    40/40

    10/10/201310/10/2013 DIPANJAN CHATTERJEE ITIS 1P97

    FALL 2010

    Southern General Hospital

    Maxim ize z =20x1 +25x2

    Subject to :

    1x1 + 1x2 10 (total bed l im itat ion )$20,000x1 + $24,600x2 $210,000 (budget con stra int)1x1 2 (at least two beds in

    materni ty ward)

    x1, x2 0 (nonnegativi tyconstra int)


Recommended