ch8_student

Embed Size (px)

Citation preview

  • 8/7/2019 ch8_student

    1/76

    Chap 8.

    The Transportation and

    Assignment Problem

    by Dr. Peitsang Wu

    Department of IndustrialEngineering and Management

    I-Shou University

  • 8/7/2019 ch8_student

    2/76

    The Transportation Problem

    m resources, n destination

    sinumber of nit supplied by source i

    djnumber of unit required by destinationj

    cijtransportation cost per unit shipped from

    source i to destinationj

  • 8/7/2019 ch8_student

    3/76

    The Transportation Problem

    Objectiveminimize total transportation cost

    xijnumber of product shipped from source i

    to destinationj

  • 8/7/2019 ch8_student

    4/76

    The Transportation Problem

    ost per unit istributed estination

    n....321 upply

    source

    m

    /

    /

    /

    /

    2

    1

    mnmmm

    n

    n

    cccc

    cccc

    cccc

    ....

    /

    /

    /

    /

    1/

    1/

    1/

    1/

    ....

    ....

    321

    2232221

    1131211

    ms

    s

    s

    /

    /

    /

    /

    2

    1

    emand ndddd ....321

  • 8/7/2019 ch8_student

    5/76

    Properties of Transportation Problem

    Feasible Solutions Property :

    If the total supply { total demand, it meaneithersi ordj represent a bound rather than anexact requirement, in this case introduce

    dummy source or dummy destination asthe slack variable.

    Integer solutions property: Not onlysi and djmust be integer values. But also all the B.F.S.

    have integer values.

    ! !! !

    !!m

    i

    n

    j

    ij

    m

    i

    n

    j

    ji xds

    1 11 1

  • 8/7/2019 ch8_student

    6/76

    Example: Production Scheduling

    Let xjbe the number of engines to be produced inmonthj.

    The problem can be formulated as the general L.P.

    model.

    MonthScheduled

    Installations

    Max

    Production

    Unit cost

    of

    production

    Unit cost

    of storage

    1 10 25 1.08 0.0152 15 35 1.11 0.015

    3 25 30 1.10 0.015

    4 20 10 1.13 0.015

  • 8/7/2019 ch8_student

    7/76

    Formulate as Transportation Problem

    si: production of jet engines in month i

    (i =1.4)

    dj: installation of jet engines in monthj(j =1.4)

    xij: number of engines produced in month i

    for installation in monthjcij: cost associated with each unit ofxij

  • 8/7/2019 ch8_student

    8/76

    Formulate as Transportation Problem

    Supply is not fixed quantities.

    Cost per unit distributed

    Destination Supply

    1 2 3 4

    source

    1

    2 ?

    3 ? ?4 ? ? ?

    Demand

  • 8/7/2019 ch8_student

    9/76

    Formulate as Transportation Problem

    In fact x11+x12+x13+x14e 25

    x21+x22+x23+x24e 35

    x31+x32+x33+x34e 30

    x41+x42+x43+x44e 10

    And notice that

    i.e. total supply {total demand)+ + + = = supply+ + + = = demand

    total supply - total demand =

    { 0demandSupply

  • 8/7/2019 ch8_student

    10/76

    Solution

    Use Big-Mmethod and introduce a new dummy

    destination d5

    Cost per unit distributed

    Destination Supply

    51 2 3 4

    source

    1 1.080 1.095 1.110 1.125 0 25

    2 M 1.110 1.125 1.140 0 353 M M 1.100 1.115 0 30

    4 M M M 1.130 0 10

    Demand 10 15 25 20 30

  • 8/7/2019 ch8_student

    11/76

    Example: Distribution of Water Resources

    Cost per Acre Foot

    Supply

    BerdooLos

    Devils

    San

    Go

    Holly

    glass

    Colombo

    River16 13 22 17 50

    Sacron River 14 13 19 15 60

    Calorie River 19 20 23 - 50

    Min needed 30 70 0 10

    requested 50 70 30 g

  • 8/7/2019 ch8_student

    12/76

    Example: Distribution of Water Resources

    Observations:

    upper bound for Holly glass:

    i.e. Holly glass can get as much as units.

    Demand: bounded variables not constant.

    regard less the minimum need, the demandis fixed.

    now

    pexcess demand.

  • 8/7/2019 ch8_student

    13/76

    1 2 3 4 Supply

    Colombo 16 13 22 17 50Sacron 14 13 19 15 60

    Calorie 19 20 23 M 50

    dummy 0 0 0 0 50demand 50 70 30 60

  • 8/7/2019 ch8_student

    14/76

    After considering the minimum needs

    1

    (min)

    2

    (extra)

    3 4 5 6Supply

    1 16 16 13 22 17 17 50

    2 14 14 13 19 15 15 60

    3 19 19 20 23 M M 50

    4 M 0 M 0 M 0 50

    demand 30 20 70 30 10 50

  • 8/7/2019 ch8_student

    15/76

    Transportation Simplex Method

    estination

    1 2 3 4 5 n Supply ui

    c11 c12 c1n

    1 s1

    c21 c22 c212 s2

    /

    / //

    /

    /

    cm1 cm2 cmn

    source

    m

    1

    1

    1

    1

    1

    sm

    emand d1 d2 d3 d4 d5 dn

    vj

    Z

  • 8/7/2019 ch8_student

    16/76

  • 8/7/2019 ch8_student

    17/76

    Procedure

    Step 1: Initialization

    Step 2: Optimality Test

    Step 3: Iteration 3

    Determine the entering basic variable

    Determine the leaving basic variable

    Determine the new B.F.S.

  • 8/7/2019 ch8_student

    18/76

    Initialization

    In general L.P. problem, well have exact onebasic variable for each constraint,For transportation problem with m sources

    and n destinations, the number of basicvariables is m+n-1.

    The reason is that the functional constraintsare equality constraints, this set ofm+n equations with one extra (or redundant)equation that can be deleted without changingthe feasible region i.e. any one of theconstraints are satisfied.

  • 8/7/2019 ch8_student

    19/76

    Procedure for Constructing the Initial B.F.S.

    1.From the rows and columns still under

    consideration, select the next basic variable

    (allocation) according to some criterion.2.Make that allocation large enough to

    exactly use up the remaining supply in its

    row or the remaining demand in its column(which ever is smaller) .

  • 8/7/2019 ch8_student

    20/76

    Procedure for Constructing the Initial B.F.S.

    3.Eliminate that row or column from furtherconsideration. (If the row and the column havethe sane remaining supply and demand, select

    arbitrary row as the one to be eliminated. Thecolumn will be use later to provide a degeneratebasic variable, i.e. a circled allocation with zero)

    4.If only one row or column remains underconsideration, then the procedure is completed byselecting every remaining variable associatedwith that row or column to be basic with the onlyfeasible allocation. Otherwise return to step 1.

  • 8/7/2019 ch8_student

    21/76

    Alternative Criteria for Initialization

    Northwest corner rule.

    Vogels approximation method.

    Russells approximation method.

  • 8/7/2019 ch8_student

    22/76

    Northwest Corner Rule

    Begin by selecting x11

    (the northwestcorner).

    Thereafter, if

    xij was the last variableselected, then next select xi , j+1 (that is,

    move one column to the right) if source ihas any supply remaining.

    Otherwise, select next xi+1 , j (that is, moveone row down) .

  • 8/7/2019 ch8_student

    23/76

    Example

    Destination

    1 2 3 4 5 Supply u i

    1 1 1 2 11 5 0

    1 1 1 1 12 6 0

    1 1 2 2 M

    3 5 0

    M 0 M 0 0

    source

    4 (D ) 5 0

    D e m a n d 3 0 2 0 7 0 3 0 6 0

    vj=

  • 8/7/2019 ch8_student

    24/76

    Vogels Approximation Method

    For each row and column remaining underconsideration, calculate its difference,which is defined as the arithmeticdifference between the smallest and thenext-to-the-smallest unit cost cij stillremaining in that row or column.

    In that row or column has the largestdifference, select the variable having thesmallest remaining unit cost. (If tie, breaksarbitrarily)

  • 8/7/2019 ch8_student

    25/76

    Example

    estination

    1 2 3 4 5 Supply di erence

    1 16 16 13 22 17 50

    2 14 14 13 19 15 60

    3 19 19 20 23 50Source

    4 0 0 0 50emand 20 70 30 60 Select x

    di erence eliminate column

  • 8/7/2019 ch8_student

    26/76

    Example

    Destination

    1 2 3 4 5 Supply difference

    1 16 16 13 17 50

    2 14 14 13 15 60

    3 19 19 20 M 50Source

    4 M 0 M 0 20

    Demand 30 20 70 60 Select x =

    difference eliminate row

  • 8/7/2019 ch8_student

    27/76

    Example

    estination

    1 2 3 5 Supply di erence

    1 16 16 13 17 50

    2 14 14 13 15 60Source

    3 19 19 20 50

    emand 30 20 70 40 Select x

    di erence eliminate ro

  • 8/7/2019 ch8_student

    28/76

    Example

    1 2 3 5 Supply difference

    2 14 14 13 15 60Source3 19 19 20 M 50

    Demand 30 20 70 40 Select x =

    difference eliminate column

  • 8/7/2019 ch8_student

    29/76

    Example

    (not column 3,use or ourth iteration in degeneracy)

    1 2 3 Supply di erence

    2 14 14 13 20S

    ource 3 19 19 20 50

    emand 30 20 20 Select x

    di erence eliminate ro

  • 8/7/2019 ch8_student

    30/76

    Example

    1 2 3 Supply

    Source 3 19 19 20 50

    emand 30 20 0

    Select x31

    x32

    x33

    z

  • 8/7/2019 ch8_student

    31/76

    Russells Approximation Method

    For each source row i remaining underconsideration, determine its, which is thelargest unit cost cij still remaining in that row.

    For each destination columnj remaining underconsideration, determine its which is thelargest unit cost cij still in that column.

    For each xij not previously selected in these rows

    and columns, calculate . Select the variable having the largest negative

    value of (Tie breaks arbitrarily).

    jv

    jiijij vuc!(

    ij(

  • 8/7/2019 ch8_student

    32/76

    Example

    Iteration1u 2u 3u 4u 1v 2v 3v 4v 5v ij( llocation

    1 x =

    2 x =

    3 x =

    4 x =

    5 x =

    6 x =

    x =

    x =

    Z =

  • 8/7/2019 ch8_student

    33/76

    Destination

    1 2 3 4 5 Supply iu

    1 16 16 13 22 17 50

    2 14 14 13 19 15 60

    3 19 19 20 23 M 50Source

    4 M 0 M 0 0 50Demand 30 20 70 30 60

    jv

  • 8/7/2019 ch8_student

    34/76

    Comparison of Three Methods

    Northwest corner : quick and easy. But itpays no attention to cij, the solution will befar from optimal.

    Vogels approximation : popular, easy toimplement by hand.

    Russells approximation : excellent

    criterion, quick implement is computer (butnot hand) better solution than the other twomethods.

  • 8/7/2019 ch8_student

    35/76

    Optimality Test

    A basic feasible solution is optimal if and

    only if for every ( i ,j )

    such that xij is non-basic.

    0u jiij vuc

  • 8/7/2019 ch8_student

    36/76

    Note

    Since cij ui vj = 0 ifxij is a basic variable, cij = ui +vj

    There are

    m+n-1 basic variables.

    m+n-1 equations,,but m+n unknowns.

    A convenient choice is to select ui that hasthe largest number of allocations in

    its row (tie broken arbitrarily), and toassign it to be zero.

  • 8/7/2019 ch8_student

    37/76

    Example

    30

    0 20

    40

    30

    30

    10

    50

    estination

    1 2 3 4 5 Supplyui

    16 16 13 22 17

    1 50

    14 14 13 19 152 60

    19 19 20 233

    -2250

    0 0 0

    Source

    4+3 +4

    50

    emand 30 20 70 30 60

    vjZ =

  • 8/7/2019 ch8_student

    38/76

    An Iteration

    STEP 1:Select the one with the largest negative value

    ofcij - ui - vj to be the entering basic variable.

    STEP 2: Increase the entering basic variable from 0 sets

    off a chain reaction of compensating changesin other basic variables.

    In order to satisfy the supply and demandconstraints. The first basic variable to bedecreased to 0 then becomes the leaving basicvariable.

  • 8/7/2019 ch8_student

    39/76

    An Iteration

    STEP 3:

    The new B.F.S. is identified by adding value

    of the leaving basic variable to the allocationfor each recipient cell and subtracting this

    same amount from the allocation for each

    donor cell.

  • 8/7/2019 ch8_student

    40/76

    Example (STEP 2)

    3 4 5 Supply

    1

    22 171 50

    1

    19 152 60

    emand 70 30 60

    Step 3: (Z=10(15-17+13-13)=10(-2)=10(c25-u2-v5)

    40

    30

    10

    +

    +

    +4

    +1

    -

    -

    -2

  • 8/7/2019 ch8_student

    41/76

    Summary

    Initialization:

    Construct an initial B.F.S. by any of the threemethods, go to optimality test.

    Optimality Test: Derives ui and vjby selecting the row having the

    largest number of allocations, setting it ui = 0 ,

    and then solving the set of equations cij = ui +vj

    for each ( i ,j ) such that xij is basic. If for every ( i ,j ) such that xij is

    non-basic, then the current solution is optimal.pSTOP.

    Otherwise, go to an iteration.

    0u jiij vuc

  • 8/7/2019 ch8_student

    42/76

  • 8/7/2019 ch8_student

    43/76

    Iteration

    STEP 3:

    Determine the new B.F.S. : Add the value of

    the leaving basic variable to the allocation foreach recipient cell. Subtract this value from

    the allocation for each donor cell. Return to

    Optimality test.

  • 8/7/2019 ch8_student

    44/76

    Transportation Simplex Tableaux

    estinationIteration

    0 1 2 3 4 5Supply ui

    16 16 13 22 171 50

    14 14 13 19 152 60

    19 19 20 233 50

    0 0 0

    Source

    4( ) 50

    emand 30 20 70 30 60

    vj

    Z =

    40 10

    30

    0 20

    30

    30

    50

  • 8/7/2019 ch8_student

    45/76

    Transportation Simplex Tableaux

    estinationIteration

    1 1 2 3 4 5Supply

    ui

    16 16 13 22 171 50

    14 14 13 19 152 60

    19 19 20 233 50

    0 0 0

    Source

    4( ) 50

    emand 30 20 70 30 60

    vjZ =

    50

    30

    0 20 30

    50

    20 10

  • 8/7/2019 ch8_student

    46/76

    Transportation Simplex Tableaux

    DestinationIteration

    2 1 2 3 4 5 Supply ui

    16 16 13 22 171 50

    14 14 13 19 152 60

    19 19 20 23 M3 50

    M 0 M 0 0

    Source

    4(D) 50

    Demand 30 20 70 30 60

    vj=

    50

    20 40

    0

    30 20

    30 20

  • 8/7/2019 ch8_student

    47/76

    Complete set of transportation simplex

    tableaux (P330,331 Table 8.23)

    DestinationIteration

    3 1 2 3 4 5 Supply ui

    16 16 13 22 171 50

    14 14 13 19 152

    +260

    19 19 20 233 50

    0 0 0

    Source

    4(D) 50

    Demand 30 20 70 30 60

    vjZ =

    50

    20 40

    30 20 0

    30 20

  • 8/7/2019 ch8_student

    48/76

  • 8/7/2019 ch8_student

    49/76

    Changing Coefficient of a Nonbasic Variable

    The change of coefficient of a non basicvariable xij will leave the r.h.s. of theoptimal tableau unchanged. Thus thecurrent basis will still be feasible.

    is not changed, ui , vj s remainunchanged.

    In row 0, only the coefficient ofxijchanged, thus as long as the coefficient ofxij in the optimal row 0 is non negativethe current basis remains optimal.

    1T

    B

    Bc

  • 8/7/2019 ch8_student

    50/76

    Assure the Optimal Tableau

    Destination

    1 2 3 4 Supply ui

    8 6 10 91

    +2 +735 0

    9 12 13 72

    +3 +250 3

    14 9 16 5

    Source

    3+5 +3

    40 3

    Demand 45 20 30 30

    vj 6 6 10 2Z =1,020

    10 25

    45 5

    10

  • 8/7/2019 ch8_student

    51/76

    Ex: Change c11

    from 8 to 8+

    Change c11

    from 8 to 8+ , for what

    values of will the current basis remains

    optimal?

    ifc11u 6 the current basis remains optimal.

    !d 1111 vuc

  • 8/7/2019 ch8_student

    52/76

    Changing Coefficient of a Basic Variable

    Since we are changing . willchange too. The coefficient of each nonbasic variable in row 0 may change.

    To determine whether the current basisremains optimal, we must find the new uis,and vj s and use these values to price out

    all non basic variables.The current basis remains optimal, as long

    as all non basic variables are non negative.

    Tc 1T

    B

    Bc

  • 8/7/2019 ch8_student

    53/76

  • 8/7/2019 ch8_student

    54/76

    Ex:Change c13

    from 10 to 10+

    ,

    ,,,

    ,

    ,,0

    010..0

    32

    1432

    4332

    2123

    31121

    313113

    !!

    !!!!

    !!

    !!!!!

    d

    !d

    d

    (!d

    uu

    vvvv

    vuvu

    vuvuvuvuulet

    vueivuc

  • 8/7/2019 ch8_student

    55/76

    Ex:Change c13

    from 10 to 10+

    Price out each non basic variable

    d

    d

    d

    d

    d

    d

    :

    :

    :

    :

    :

    :

    33

    31

    24

    22

    14

    11

    c

    c

    c

    c

    c

    c

  • 8/7/2019 ch8_student

    56/76

    Increasing bothsi and djby

    This change maintains a balance transportation

    problem.

    Since bothsi

    and dj

    increase, this will cancel

    out the supply demand balance.

    Now.

    New value = old -value +ui +vj

    e.g. increase source 1 and destination 2 by 1 unit.

    New cost = 1020+1*0+1*(6) = 1026

  • 8/7/2019 ch8_student

    57/76

    Increasing bothsi and djby

    We may also find new values of the decisionvariables as follows:

    (i) Ifxij is a basic variable in the optimal solution,

    increase xijby(ii) Ifxij is a non basic variable in the optimal

    solution, find the loop involving xij and someof the basic variables. Find an odd cell in the

    loop that is in row i. Increase the value of thisodd cell by and go around the loop,alternatively increasing and then decreasingcurrent basic variables in the loop by.

  • 8/7/2019 ch8_student

    58/76

  • 8/7/2019 ch8_student

    59/76

    Ex: Increases1, d

    2by 2

    x12

    is a basic variable in the optimal solution.

    Now

  • 8/7/2019 ch8_student

    60/76

    Ex: Increases1, d

    1by 1

    x11 is a non basic variable in the optimal solution.

    Destination

    1 2 3 4 Supply ui

    1 35 0

    2 50 3Source

    3 40 3

    Demand 45 20 30 30

    vj 6 6 10 2

    Z=1020+u1+v1

    =

    10 26

    46 4

    10 30

  • 8/7/2019 ch8_student

    61/76

    The Assignment Problem

    Assumptions:

    The number of assignees and the number of tasksare the same. (denote by n)

    Each assignee is to be assigned to exactly on task.

    Each task is to be performed by exactly oneassignee.

    There is a cost cij

    associated with assignee iperforming taskj ( i ,j = 1..n )

    The objective is to determine how all nassignments should be made in order to minimizethe total cost.

  • 8/7/2019 ch8_student

    62/76

    Example

    ocation

    n =4

    ocation

    n =4

    1 2 3 4 1 2 3 41 13161211 1 13161211

    2 15 - 1320 2 15 1220

    achine

    n =3

    3 5 7 10 6 3 5 7 10 6

    achine

    n =4

    4

    (D)0 0 0 0

  • 8/7/2019 ch8_student

    63/76

    The Mathematical Model

    ( actually xij n_a )

    jix

    x

    xts

    xcZMin

    ij

    n

    i

    ij

    n

    j

    ij

    n

    j

    n

    i

    ijij

    ,,0

    1

    1..

    1

    1

    1 1

    u

    !

    !

    !

    !

    !

    ! !

    .

    .0

    1..

    jtaskperformsi

    not

    assignee

    if

    ifxei ij

    !

  • 8/7/2019 ch8_student

    64/76

    The Mathematical Model

    Assignment problem is a special case of

    transportation problem:

    Number of sources m = number ofdestination n

    Every supply si = 1

    Every demand dj = 1

  • 8/7/2019 ch8_student

    65/76

    The Two Properties

    Feasible solution property

    Integer solution property holds here.

    More over, the constraints prevent anyvariable from being greater than 1, and the

    non-negativity constraints prevent values

    less than 0. So it automatically satisfiesbinary restriction.

  • 8/7/2019 ch8_student

    66/76

  • 8/7/2019 ch8_student

    67/76

    Our example

    1 2 3 4 Supply

    1 13 16 12 11 1

    2 15 M 13 20 1

    3 5 7 10 6 1

    4(D) 0 0 0 0 1

    Demand 1 1 1 1

  • 8/7/2019 ch8_student

    68/76

  • 8/7/2019 ch8_student

    69/76

  • 8/7/2019 ch8_student

    70/76

    Hungarian Method:

    STEP 2:

    Draw the minimum number of lines

    (horizontal and/or vertical) that are needed to

    cover all the zeros in the reduced cost matrix.

    Ifn lines are required to cover all the zeros, an

    optimal solution is available among the cover

    zeros in the matrix. If fewer than n lines are needed to cover all the

    zeros, GOTO STEP 3.

  • 8/7/2019 ch8_student

    71/76

    Hungarian Method:

    STEP 3:

    Find the smallest nonzero element (call itsvalue k) in the reduced cost matrix that is not

    covered by lines draw n in STEP2.Now subtract kfrom each uncovered element

    of the reduced cost matrix and add kto eachelement of the reduced cost matrix that is

    covered by two lines.Return to STEP 2.

  • 8/7/2019 ch8_student

    72/76

    Example

    row min

    14 5 8 7

    2 12 6 5

    7 8 3 9

    2 4 6 10

    Column

    min

  • 8/7/2019 ch8_student

    73/76

    Example

    9 0 3 0

    0 10 4 1 p

    4 5 0 4Smallest

    Uncovered:1

    0 2 4 6

  • 8/7/2019 ch8_student

    74/76

    Transshipment Problem

    A supply node is a node that can send goods to

    another node but cannot receive goods from any

    other node.

    A demand node is a node that can receive goods

    from other nodes but cannot send goods to any

    other node.

    A transshipment node is a node that can bothreceive goods from other nodes and send goods

    to other nodes.

  • 8/7/2019 ch8_student

    75/76

    Solution Procedure

    Step 1: If necessary, add a dummy demand tobalance the problem.s =total available supply.

    Step 2: Construct a transportation tableau as

    follow:A row consists of supply node and

    transshipment node.

    A column consists of demand node and

    transshipment node.The transshipment node will have a supply

    equal to (nodes original supply)+s, and ademand equal to (nodes original demand)+s.

  • 8/7/2019 ch8_student

    76/76

    Example