Transportation problem The transportation problem is a special class of the LPP.

It deals with the situation in which a commodity is shipped from sources( e.g., factories) to destinations (e.g., warehouses). The objective is to determine the amounts shipped from each source to each destination that minimize the total shipping cost while satisfying both the supply limits and the demand requirements. The model assumes that the shipping cost on a given route is directly proportional to the number of units shipped on that route. The transportation problem applies to situations in which a single product is to be transferred from several sources to several sinks. In general, let there be m sources S1, S2, Sm having ai(I =1,2,,m_ units of supplies or capacity respectively to be transported among n destinations, D1, D2,.. Dn with bj(j=1,2,,n) units of requirements respectively. Let cij be the cost of shipping one unit of the commodity from sources I to destination j for each route. If xij represents the units shipped per route from source I to destination j, then the problem is to determine the transportation schedule so as to minimize the total transportation cost satisfying supply and demand conditions. Mathematically, the problem may be stated as follows:
m n

Minimize (total cost) Z = cij xij

i=1 j=1

Subject to the constraints

n

xij = ai, i=1,2,.,m (supply constraint)

j=1 m

xij = bj, j =1,2,.,n (demand constraint)

i=1

and xij >= 0 for ease in presentation and solution, a transportation problem is generally presented as shown in following table:
...

To From S1 S2 : Sm

D1

D2 .. .... : .

Dn

Supply ai

C11 x11 C21 x21 : Cm1 xm1

C12 x12 C22 x22 : Cm2 xm2

C1n x1n C2n x2n : Cmn xmn

a1 a2 : am

Demand bj

b1

b2

..

bn

ai, = bj

ai, = bj demand.
i=1 j=1

means total supply is equal to total

If the above condition is satisfied then the Transportation Problem is balanced. Any set of non-negative allocations (xij > 0), which satisfies the row and column sum, is called a feasible solution. A Feasible solution is known as Basic Feasible Solution if the number of non-negative allocations is equal to m + n 1 where m is number of rows and n is number of columns in a TT. Any Feasible Solution to a Transportation Problem containing m origins and n destinations is said to be non-degenerate, if it contains m + n 1 occupied cells and each allocation is in independent positions. If a basic solution contains less than m + n 1 non-negative allocations, it is said to be degenerate.

A feasible solution, which minimizes the total cost is known as Optimum solution.

Finding of Initial Basic Feasible Solution The solution of a transportation problem can be obtained in two stages, namely initial solution and optimum solution. Initial solution can be obtained by using any one of the following three methods: North West Corner Rule Least Cost method (LCM)/ Matrix Minima Method Vogels Approximation Method (VAM)/ Penalty Method VAM is preferred over the other two methods, since the initial basic feasible solution obtained by this method is either optimum or very close to the optimum solution. The improved solution of the initial basic solution is called optimal solution which is the second stage of solution, that can be obtained by Stepping Stone Method or MODI (Modified Distribution Method) Method NWCM Step1: Starting with the cell at the upper left corner (North West) of the transportation matrix we allocate as

much as possible so that either the capacity of the first row is exhausted or the destination requirement of the first column is satisfied i.e., x11 = min (a1, b1). Step 2. If b1 > a1, we move down vertically to the second row and make the second allocation of magnitude x22 = min (a2, b1-x11) in the call (2,1). If b1 < a1, move right horizontally to the second column and make the second allocation of magnitude x12 = min (a1, x11 b1) in the cell (1,2). If b1 = a1 there is a tie for the second allocation. We make the second allocation of magnitude X12 = min (a1 a1, b1) = 0 in cell (1,2) Or x21 = min (a2, b1 b1) = 0 in the cell (2,1) Step 3. Repeat steps 1 and 2 moving down towards the lower right corner of the transportation table until all the requirements are satisfied. Example: Obtain the initial basic feasible solution of a transportation problem whose cost and requirement table is given below Origin\Destination O1 O2 D1 2 3 D2 7 3 D3 4 1 Supply 5 8

5 4 7 7 1 6 2 14 7 9 18 34 Table 1 Solution: Given Transportation Problem is balanced, as total supply is equal to total demand. The first allocation is made in cell (1,1) the magnitude being x11 = min (5,7) = 5. Origin\Destination O1 2 5 O2 3 O3 5 O4 1 Demand 7 6 9 2 18 4 7 14 3 1 7 D1 7 D2 4 8 D3 Supply 5

O3 O4 Demand

The second allocation is made in the cell (2,1) and the magnitude of the allocation is given by x21 = min(8, 7-5) = 2.

Origin\Destination O2 3 2 O3 5 O4 1 Demand

D1 3 4 6 2

D2 1

D3

Supply 8 7

7 14 2 9 18

The third allocation is made in the cell (2,2) and the magnitude of the allocation is given by x22 = min(8-2, 9) = 6. Origin\Destination O2 3 6 O3 4 O4 6 2 7 14 D2 1 7 D3 Supply 6

Demand

18

The fourth allocation is made in the cell (3,2) and the magnitude of the allocation is given by x32 = min(7, 9-6) = 3. Origin\Destination O3 4 3 O4 6 Demand 3 2 18 D2 7 14 D3 Supply 7

The fifth allocation is made in the cell (3,3) and the magnitude of the allocation is given by x33 = min(7-3, 14) = 4. Origin\Destination O3 7 4 O4 2 14 D3 Supply 4

Demand

18

The final allocation is made in the cell (4,3) and the magnitude of the allocation is given by x43 = min (14,18-4) = 14. Origin\Destination O4 Demand D3 2 14 14 Supply 14

Hence we get the initial basic feasible solution to the given Transportation Problem and is given by x11 = 5, x21 = 2, x22 = 6, x32 = 3, x33 = 4, x43 = 14 Total Cost = 2 x 5 + 3 x 2 + 3 x 6 + 3 x 4 + 4 x 7 + 2 x 14 = 10 + 6 + 18 + 12 + 28 + 28 = Rs. 102. Example: Determine an initial basic feasible solution to the following Transportation Problem using N.W.C.R Origin\Destination O1 O2 O3 D1 6 8 4 D2 4 9 3 D3 1 2 6 D4 5 7 2 Supply 14 16 5

Demand

10

15

35

Solution: The problem is balanced Transportation Problem as the total supply is equal to the total demand. Using the steps we find the initial basic feasible solution. Origin\Destination O1 6 6 O2 8 O3 4 Demand 6 3 10 6 1 15 2 4 4 35 9 2 2 14 7 5 D1 4 8 D2 1 D3 5 16 D4 Supply 14

Hence we get the initial basic feasible solution to the given Transportation Problem and is given by x11 = 6, x12 = 8, x22 = 2, x23 = 14, x33 = 1, x34 = 4 Total Cost = 6 x 6 + 4 x 8 + 2 x 9 + 2 x 14 + 6 x 1 + 2 x4 = 10 + 6 + 18 + 12 + 28 + 28 = Rs. 128 Least Cost Method

Step1. Determine the smallest cost in the cost matrix of the TT. Let it be cij. Allocate xij = min (ai, bj) in the cell (i,j). Step 2. if xij = ai cross off the ith row of the TT and decrease bj by ai. Then go to step 3. If xij = bj cross off the jth column of the TT and decrease ai by bj. Then go to step 3. If xij = ai = bj cross off both ith row & the jth column. Step 3. Repeat steps 1 & 2 for the resulting reduced TT until all the requirements are satisfied. Whenever the minimum cost is not unique, make an arbitrary choice among the minima. Example: Obtain an initial basic feasible solution to the following Transportation Problem using LCM. Origin\Destination O1 O2 O3 Demand D1 1 4 0 4 D2 2 3 2 6 D3 3 2 2 8 D4 4 0 1 6 Supply 6 8 10 24

Solution: Since ai, = bj = 24, there exists a feasible solution to the Transportation Problem using the steps in the

LCM, the first allocation is made in the cell (3,1) the magnitude being x31 = 4. which satisfies the demand at the destination Di and we delete this column from the table as it is exhausted. Origin\Destination O1 1 O2 4 O3 0 4 Demand 4 The remaining table is Origin\Destination O1 2 O2 3 O3 2 2 1 2 0 6 6 D2 3 2 6 D3 4 8 2 8 D4 1 6 Supply 6 3 2 0 10 D1 2 D2 3 D3 4 8 D4 Supply 6

Demand 6 8 6 The second allocation is made in the cell (2,4) with magnitude x24 = 6. since it satisfies the demand at the destination D4, it is deleted from the table. The residual table is shown below. Next allocation is made as x12 = 6 and both O1 row and D2 column will be deleted. Origin\Destination D2 D3 Supply O1 6 2 3 6 O2 2 3 2 O3 2 Demand 6 2 8 6

After the deletion of O1 row and D4 column we get the residual table as shown below. Next allocations are x23 = 2 and x33 = 6. Origin\Destination O2 2 D3 Supply 2

2 O3 2 6 Demand 8 Hence all the allocations can be shown in the following table: Origin\Destination O1 1 O2 4 O3 0 4 Demand 4 2 6 2 6 8 1 6 3 2 2 0 6 10 D1 2 6 D2 3 D3 4 8 D4 Supply 6 6

the initial basic feasible solution to the given Transportation Problem and is given by x12 = 6, x23 = 2, x24 = 6, x31 = 4, x33 = 6 Total Cost = 6 x 2 + 2 x 2 + 6 x 0 + 4 x 0 + 6 x 2 = 12+ 4 + 12 = Rs. 28

The solution of the given Transportation Problem is degenerate as number of allocations are less than n + m1 Example: Determine an initial basic feasible solution to the following Transportation Problem using LCM. Origin\Destination O1 O2 O3 Demand D1 6 8 4 6 D2 4 9 3 10 D3 1 2 6 15 D4 5 7 2 4 Supply 14 16 5 35

Solution: The problem is balanced Transportation Problem as the total supply is equal to the total demand. Using the steps of LCM we find the initial basic feasible solution. Origin\Destination O1 6 O2 8 O3 4 3 6 2 9 2 7 5 D1 4 D2 D3 1 14 5 16 D4 Supply 14

Demand 6 10 15 4 Table 1 Minimum cost exists in the cell x13, so the first allocation is made as x13 = 14 as shown in table 1 and residual table is obtained by deleting O1 row as shown in table 2.

Origin\Destination O2 8 O3 4

D1 9 3

D2 2 6

D3 7

D4

Supply 16 5

2 4

Demand 6 10 1 4 Table 2 Now minimum cost is 2 in both the cells (2,3) & (3,4), we select (3,4) cell and made allocation as x34 = 4 shown in table 2. Residual table is obtained by deleting D4 column as shown in table 3. Origin\Destination O2 8 D1 9 D2 2 D3 Supply 16

1 O3 4 3 6 1

Demand 6 10 1 Table 3 Now minimum cost is 2 in the cell (2,3) and allocation is done as x23 = 1 shown in table 3. Residual table is obtained by deleting D3 column as shown in table 4. Origin\Destination O2 8 O3 4 3 1 D1 9 1 D2 Supply 15

Demand 6 10 Table 4 Now minimum cost is 3 in the cell (3,2) and allocation is done as x32 = 1 shown in table 4. Residual table is obtained by deleting O3 row as shown in table 5. Origin\Destination O2 D1 D2 Supply 15

8 6 Demand Table 5 6

9 9 9

Now single row is remaining so there is no choice left. So allocations are x21 = 6 & x22 = 9 as shown in table 5 and overall allocations are shown in table 6.

Origin\Destination O1 6 O2 8 6 O3 4 Demand Table 6

D1 4 9 9 3 1 6

D2

D3 1 14 2 1 6 5

D4

Supply 14 16

7 5 2 4 15 4

10

Hence we get the initial basic feasible solution to the given Transportation Problem and is given by x13 = 14, x21 = 6, x22 = 9, x23 = 1, x32 = 1, x34 = 4 Total Cost = 14 x 1 + 6 x 8 + 9 x 9 + 1 x 2 + 1 x 3 +4 x2 = 14 + 48 + 81 + 2 + 3 + 8 = Rs. 156. Vogels Approximation Method Step1. Find the penalty cost, namely the difference between the smallest and next smallest costs in each row and column. Step 2. Among the penalties as found in step 1 choose the maximum penalty is more than one (in case of tie) choose any one arbitrary. Step 3. In the selected row or column as by step 2. Find out the cell having the least cost. Allocate to this cell as much as possible depending on the capacity and requirements. If the column is exhausted, then there is a change in row penalty and vice versa. Example: Determine an initial basic feasible solution to the following Transportation Problem using LCM. Origin\Destination D1 D2 D3 D4 Supply

O1 O2 O3 Demand

11 16 21 200

13 18 24 225

17 14 13 275

14 10 10 250

250 300 400 950

Solution: Since ai, = bj = 24, there exists a feasible solution to the Transportation Problem. First we find the row and column penalty P1 as the difference between the least cost and next least cost. The maximum penalty is 5. Choose the first column arbitrarily. In this column choose the cell having the least cost name (1,1). Allocate to this cell with minimum magnitude (i.e., (250,200) = 200). This exhausts the first column. Delete this column. Since column is deleted, then there is a change in row penalty P11 and column penalty P11 remains same. Continuing in this manner we get the remaining allocations as given in the table below. Origin\Destination D1 D2 D3 D4 Supply PI O1 250 2 11 13 17 14 200 O2 300 4 16 18 14 10 O3 400 3

21 Demand PI I allocation Origin\Destination O1 O2 18 O3 24 Demand PII II allocation Origin\Destination O2 O3 24 225 5

24

13

10 950

200 225 275 250 5 5 3 0 D2 13 50 D3 17 14 13 275 3 D4 14

Supply 50 300

PII 3 4 3

10 400 10 250 0 950

D2 18 175

D3 14 13

D4 10

Supply PIII 300 4 400 3

10

Demand PIII III allocation Origin\Destination O2 O3

175 6 D3 14 13

275 1 D4 10 125

250 0 Supply 125 400

950

PIV 4 3

10 250 0 950

Demand PIV IV allocation

275 1

Origin\Destination O3 Demand V & VI allocations

D3

D4

Supply 400 950

13 10 275 400 275 125

Finally we arrive at the initial basic feasible solution, which is shown in the following table.

Origin\Destination O1 O2

D1 11 200 16

D2 13 50 18 175 24

D3 17 14

D4 14

Supply 250 300

10 125 400

O3 21 Demand 200 13 10 275 125 225 275 250

950

There are 6 positive independent allocations which equals to m + n 1. This requires that the solution is a non-degenerate basic feasible solution. Hence we get the initial basic feasible solution to the given Transportation Problem and is given by x11 = 200, x12 = 50, x22 = 18, x24 = 125, x33 = 275, x34 = 125 Total Cost = 200 x 11 + 50 x 13 + 18 x 175 + 125 x 10 + 275 x 13 +125 x 10 = 12,075 To Find Optimum Solution Once initial solution has been found, the next step is to test that solution for optimality. The following two methods are widely used for testing the solution.

Modified Distribution (MODI) Method. The modified distribution method has a pattern similar to that of the stepping stone method except that it evaluate each of the unoccupied cells in the process more efficiently. In the stepping stone method a closed path is traced for each of the unoccupied cells before the savings in respective costs, if any, can be calculated. The unoccupied cell with the largest saving (identified as the one with the most negative value) enters in the next solution. In the modified distribution method, however, the improvement (opportunity) costs of all the unoccupied cells are calculated without having to trace their respective closed paths. In fact, we need to trace only one closed path in the modified distribution method after the unoccupied cells with the most negative value has been identified. Thus, it can often provide considerable time saving over the stepping stone method for solving transportation problems.
1.

MODI Algorithm. The mechanics of the modified distribution method for finding out the optimal solution of a transportation problem involves the following steps:STEPS 1:- Determine an initial basic feasible solution consisting of m+n-1 allocations in

independent position using any of the three methods discussed earlier. 2. Determine a set of numbers for each row and each column. To compute ui (i=1,2,.,m) for each row and vj (j=1,2,..,n) for each column, set cij = ui+vj for each of the m+n-1 occupied cells used in the initial solution. Since there are m+n-1 occupied cells used in the initial solution, consequently there are m+n-1 equations involving cij, ui and vj. Because the number of unknowns ui and vj are m+n-1, we can assign an arbitrary value (conveniently zero) to one of the variables without violating the equations. 3. Compute the opportunity cost (improvement index) by using the relationship: ij = cij (ui+vj) ; (i=1,2,,m and j=1,2, ,n) For each of the unoccupied (empty) cells. In other words, for every unoccupied cell we determine its implied cost* by adding the corresponding row number and column number and then the opportunity cost is obtained by subtracting the actual cost of this cell from its implied cost. 4. Check the sign of each opportunity cost. If the opportunity costs of all the unoccupied cells are either positive or zero, the given solution is the optimal one. On the other hand, if one or more unoccupied cells have negative opportunity cost,

the given solution is not an optimal one and further saving in transportation cost are possible. 5. select the unoccupied cells with the largest opportunity cost as the cell to be included in the next solution. 6. draw a closed path or loop for the unoccupied cell selected in step 5. it may be noted that right angle turns in this path are permitted only at occupied cells at the original unoccupied cells. It may be noted that every closed loop have even numbers of turns and is found with horizontal and vertical lines only. 7. assign alternate plus and minus signs at the unoccupied cells on the corner points of the closed path with a plus sign at the cell being evaluated. 8. determine the maximum number of units that should be shipped to this unoccupied cells. The smallest stone with a negative position on the closed path indicates the numbers of units that can be shift to the entering cell. This quantity is added to all the cells marked with plus sign and is subtracted from those cells marked with minus sign in this way the unoccupied cell under consideration becomes an occupied cells making one of the occupied cells as an unoccupied one. 9. repeat the whole procedure until an optimal solution is attained, i.e.,we are able to show that

for a given solution, the opportunity cost of all the unoccupied cells are either positive or zero. * the implied cost (for a given solution) for an unoccupied cell may be understood as an upper limit on the actual shipping cost of this cell beyond which the inclusion of this cell will not decrease the transportation cost. 10. finally, obtain the total transportation cost for the new solution. Example 1:In order to demonstrate the MODI Method, let us take initial solution obtained by VAM as shown in table 1. To illustrate the MODI method. We have to alter the transportation table by assigning an additional row and column. The modified transportation table of the initial solution is shown in the table 2.

Origin\Destination W1 W2 W3 W4 F1 21 16 25 13 11 F2 17 18 14 23 6 3 4 F3

Supply 11 13 19

32 Demand Table 1 6

27 7 10

18 12 12

41 15 43

Origin\Destination W1 W2 W3 W4 Supply F1 21 16 25 13 11 F2 17 18 14 23 6 3 4 F3 32 27 7 6 10 v1 v2 18 41 12 12 15 v3 v4 19 13 11

Row Number(ui) u1 u2 u3

Demand 43 Column Number(vj) Table 2 In Table 2, since the number of occupied cells are m + n 1 = 3 + 4 1 = 6, the initial solution is nondegenerate. Thus, an optimal solution can be obtained. Steps 1: Note that we have added column ui to indicate row values and row vj to indicate column

values. Now let us define these variables before we proceed further. ui = Value for the ith row, i = 1,2, 3 vj = Value for the jth column, j = 1,2, 3,4 For the occupied cells, the following relationship exists: cij = ui + vj For example, the unit TC for the six occupied cells can be described as: c14 = u1 + v4 = 13, c21 = u2 + v1 = 17, c22 = u2 + v2 = 18, c24 = u2 + v4 = 23, c32 = u3 + v2 = 27, c33 = u3 + v3 = 18. In the above equations, we have seven unknown variables and six equations. In order to obtain a solution to determine the row and column numbers, one of the variables must be chosen and given an arbitrary value of zero. We select u2 and assign a zero value to it. With u2 = 0, we can identify the values of the remaining variables in the above relationships which are shown below: u2 + v1 = 17 0 + v1 = 17 v1 = 17 u2 + v2 = 13 0 + v2 = 18 v2 = 18 u2 + v4 = 13 0 + v4 = 23 v4 = 23 u1 + v4 = 13 u1 + 23 = 13 u1 = -10 u3 + v2 = 27 u3 + 18 = 27 u3 = 9 u1 + v4 = 13 9 + v3 = 18 v3 = 9

2. We now proceed to evaluate cost change for all the UC cells by using the formula: ij = cij (ui + vj) As in the stepping stone method, if an UCC has a negative cost change, it indicates that an improved solution is possible. When all cost changes have zero or positive values, the optimal solution is reached. The net cost change (or opportunity cost) for each of the UCC are evaluated as shown in the following table: UCC ij = cij (ui + vj) Net Cost change (F1, W1) 11 = 21 ( -10 +14 + 17) (F1, W2) 12 = 16 ( -10 +8 + 18) (F1, W3) 13 = 25 ( -10 +26 + 9) (F2, W3) 23 = 14 ( 0 + +5 9) (F3, W1) 31 = 32 ( 9 + +6 17) (F3, W4) 34 = 41 ( 9 + +9 23) The following table represents the solution of the problem with the corresponding row and column numbers.

Origin\Destination W1 W2 W3 W4 Supply F1 21 16 25 13 14 8 26 11 F2 17 18 14 23 6 3 5 4 F3 Demand Column Number(vj) Table 3 32 27 18 41 6 7 12 9 6 10 12 15 v1 v2 v3 v4 19 43 13 11

Row Number(ui) u1 u2 u3

The net cost change derived above corresponds with those we calculated in stepping stone method. The cost change calculated for all the unoccupied cells are also shown in the left hand side bottom of the unoccupied cells and these values of cost change indicate that there is no negative value. According to the optimality criteria for minimization transportation problem the current solution is optimal one, since the opportunity cost of the unoccupied cells are all zero or positive. The optimal solution to the given problem is ,then

x14=11,x21=6,x22=3,x24=4,x32=7,x33=12 and all other xij=0. The total transportation cost is Rs 796. Example 2: A company is spending rs. 1000 on T of its units from three plants to four distribution centers. The supply and demand of units, with unity cost of T are given as: Plant/Distribution Centres P1 P2 P3 Requirements Table 1 D1 19 70 40 5 D2 30 30 10 8 D3 50 40 60 7 D4 Availability 12 60 20 15 7 10 18

What can be the maximum saving by optimal schedule? Solution: Steps 1: Using VAM, an initial feasible solution of the above problem is shown in following table. Plant/Distribution D1 D2 D3 D4 Availability Centres P1 19 30 50 12 7

5 P2 P3 Requirements Table 2 70 40 8 5 8 7 30 7 10 60 40

2 60 3 20 10 15 18 10

The TC according to the above situation comes out to be (5 x 19) + (2 x 12) + (7 x 40) + (3 x 60) + (8 x 10) + ( 20 x 20) = Rs. 859. Further, the solution is basic feasible solution one as there are 4 + 3 1 = 6 allocations in independent positions, i.e., initial solution is non-degenerate as number of OC are 6. Step 2: determine ui and vj u1 + v1 = c11 0 + v1 = 19 v1 = 19 u1 + v4 = c14 0 + v4 = 12 v4 = 12 u2 + v3 = c23 48 + v3 = 40 v3 = -8 u2 + v4 = 60 u2 + 12 = 60 u2 = 48 u3 + v2 = 10 8 + v2 = 10 v2 = 2 u3 + v4 = 20 v3 + 12= 20 v3 = 8 Step 3: Evaluate the opportunity cost for each UCC by using the relationshipij = cij (ui + vj).

UCC (P1, D2) (P1, D3 (P2, D1) (P2, D2) (P3, D1) (P3, D3)

ij = cij (ui + vj) Net Cost change +28 12 = 30 (0 + 2) +58 13= 50 ( 0 - 8) +3 21 = 70 ( 48 + 19) -20 22 = 30 ( 48 + 2) +13 31 = 40 (8 + 19) +60 33 = 60 ( 8 8)

Table 3 Since the OC in the cell (P2, D2) is negative, the current basic feasible solution is not optimal and can be further improved towards optimality. Plant/Distributio n Centres P1 P2 D1 D2 D3 D4 Availability 19 5 70 3 30 28 30 50 58 40 7 12 2 60 3 7 10 Row Number(ui) u1 = 0 u2 = 48

P3 Requirements Column number(vj)

20 + 40 10 60 20 13 8 60 10 + 5 8 7 15 v1 v2 v3= v4 = =2 -8 = 19 12

18 35

u3 = 8

Table 4 Since the cell (P2, D2) has the negative opportunity cost 20, it is admitted as an entering variable in the solution Result is shown in table 5. Plant/Distribution D1 D2 D3 Centres P1 19 30 50 5 P2 70 30 40 3 7 P3 40 10 60 5 Requirements 5 8 7 Table 5 D4 Availability 12 2 60 20 13 15 10 18 7

u1 + v1 = 19, u1 + v4 = 12, u2 + v3 = 30, u3 + v2 = 40, u3 + v2 = 10, u3 + v4 = 20 Setting u1 =0 as before v1 = 19, v4 = 12, u3 = 8, v2 = 2, u2 = 28, v3 = 12. Step 3: Evaluate the opportunity cost for each UCC by using the relationshipij = cij (ui + vj). UCC (P1, D2) (P1, D3 (P2, D1) (P2, D4) (P3, D1) (P3, D3) Table 6 Plant/Distributio n Centres D1 D2 D3 D4 Availability Row Number(ui) ij = cij (ui + vj) Net Cost change +28 12 = c12 (u1 + v2) +38 13= c13 (u1 + v3) +23 21 = c21 (u2 + v1) +20 24 = c24 (u2 + v4) +13 31 = c31 (u3 + v1) +40 33 = c33 (u3 + v3)

P1 P2 P3 Requirements Column number(vj)

19 30 50 5 70 30 40 3 7 40 10 60 3 5 8 7

12 2 60 20 13 15

7 10 18 35

u1 = 0 u2 = 28 u3 = 8

v1 v2 v3= v4 = =2 12 = 19 12

Table 7 The TC according to the above situation comes out to be (5 x 19) + (2 x 12) + (3 x 30) + (7 x 40) + (5 x 10) + ( 13 x 20) = Rs. 799. 2. Stepping Stone Method In this method, we calculate the net cost change that can be obtained by introducing any of the unoccupied cells into the solution. The important rule to keep in mind is that every increase in supply at occupied cell must be associated by a decrease in supply at another. The same rule holds for demand. Steps included in stepping stone method are: Step 1: Determine an initial basic feasible solution using any of the three methods discussed earlier.

Step 2: Make sure that the number of occupied cells is exactly equal to m+n-1, where m is number of rows and n is number of columns. Step 3: Evaluate the cost effectiveness of shipping goods via transportation routes not currently in the solution. The following five steps conduct this testing of each unoccupied cell as follows: a. Select an unoccupied cell, where a shipment should be made. b. Beginning at this cell, trace a closed path using the most direct route through at least three occupied cells used in the solution and then back to original occupied cell and moving with only vertical and horizontal moves. Further, since only the cell at the turning points are considered to be on the closed path, both unoccupied and occupied boxes may be skipped over. The cells at the turning points are called stepping stones on the path. c. Assuming plus(+) and minus(-) signs on alternative corners of the closed path just traced, starting with the unoccupied cell. d. Compute the net change in the cost along the closed path by adding together the unit cost figures found in each cell containing a plus sign and then subtracting the unit costs in each square containing the minus sign.

e. Repeat sub-steps (a) and (b) to calculate net change in the cost of all unoccupied cells. Step 4: Check the sign of each of the net changes. If all net changes computed are greater than or equal to zero, an optimum solution has been reached. If not, it is possible to improve the current solution and decrease total shipping costs. Step 5: Select the unoccupied cells having the highest negative net cost change and determine the maximum number of units that can be assigned to a cell marked with a minus sign on the closed path corresponding to this cell. Add this number to unoccupied cell and to all other cells on the path marked with plus sign. Subtract this number from the cells on the closed path marked with a minus sign. Step 6: Go to step 2 and repeat the procedure until we get an optimum solution. Example: Consider following problem with initial basic solution: Origin\Destination O1 O2 D1 21 D2 16 D3 25 D4 13 11 13 Supply 11

17 1 O3 Demand Table 1 32 5 6

18 27 10 10

14 12 18 12

23 19 41 4 15 43

Step 1: The initial solution has 4+3-1 = 6 occupied cells and involves cost of Rs. 922. Step 2: Let us evaluate the net cost change for unoccupied cell (O1, D1). The shipment of one unit to this cell caused an additional cost of Rs. 21. this inturn requires that one unit should be decreased from the cell (O1, D4) which decreases cost by Rs. 13. but to keep trhe balance between capacity and the requirement we have to add one unit to the cell (O3, D4) which increases the cost by Rs. 41 and finally one unit is decreased from cell (O3, D1) which decreases cost by Rs. 32. to determine the net cost change, let us list down the changes as shown below:

Cell 13

Change in Allocation

Cost Change(Rs)

21

+ (O1, D1) 41 (O1, D4) + (O3, D4) 4 (O3, D1) +1 -1 +1 -1 -32 +17 +21 -13 +41 -

11

32

Net Cost Change

This indicates that if the occupied cell (O1, D1) is made occupied, then the total transportation cost will be increased by Rs. 17, per unit supplied. This transfer of shipment one unit is also shown in the right hand side table by making a close path. Similarly other unoccupied cells can also be evaluated proceeding in the same manner. Unoccupied Closed Path Net Cost Change Remarks Cells (O1, D2) (O1, D2)-> (O1, D4) 16-13+41-27=17 Cost Increases (O3, D4)-> (O3, D2) (O1, D3) (O1, D3)-> (O1, D4) 25-13+41-32+1714=24 Cost Increases (O3, D4)-> (O3, D1)

(O2, D1)-> (O2, D3) (O2, D2) (O2, D2)-> (O3, D2) Cost Increases (O3, D1)-> (O2, D1) (O2, D4) (O2, D4)-> (O3, D4) Cost Decreases (O3, D1)-> (O2, D3) (O3, D3) (O3, D3)-> (O3, D1) Cost Decreases (O2, D1)-> (O2, D3)

18-27+32-17=6 23-41+32-17=-3 18-32+17-14=-11

Above table shows allocation in (O3, D3) will decreases the overall cost. Therefore, the unoccupied cell (O3, D3) will be considered for further reduction in the cost. Now it is to be decided how much quantity to be allocated in this cell. The maximum quantity that can be shipped to cell (O3, D3) is exactly the minimum quantity to those cells with the minus sign in the closed path as shown in the table 2. In this case cell (O3, D1) has 5 and cell (O2, D3) has 12 as the quantity to be shipped. Therefore 5 is the mnaximum quantity that can be shipped in the cell (O3, D3) and after doing maximum allocations in the transportation table we get table 3. Origin\Destination D1 D2 D3 D4 Supply O1 11 21 16 25 13 11

O2 17 1 + O3 32 5Demand Table 2 Origin\Destination O1 O2 17 6 O3 32 Demand Table 3 6 27 10 10 18 5 12 41 4 15 18 14 7 23 6 D1 21 27 10 10 D2 16 18 + 12 D3 25 41 4 15 D4 13 11 18 14 23 12 -

13

19

43 Supply 11 13 19 43

Now Transportation Cost is 867 Rs. Next Iteration Of Stepping Stone Method.

Unoccupied Closed Path Remarks Cells (O1, D1) (O1, D1)-> (O1, D4)-> 17=28 Cost Increases (O3, D4)-> (O3, D3)-> (O2, D3)-> (O2, D1) (O1, D2) (O1, D2)-> (O1, D4)-> Cost Increases (O3, D4)-> (O3, D2) (O1, D3) (O1, D3)-> (O1, D4)-> Cost Increases (O3, D4)-> (O3, D3) (O2, D2) (O2, D2)-> (O2, D3)-> Cost Decreases (O3, D3)-> (O3, D2) (O2, D4) (O2, D4)-> (O3, D4)-> Cost Decreases (O3, D3)-> (O2, D3) (O3, D1) (O3, D1)-> (O3, D3)-> Cost Increases (O2, D3)-> (O2, D1)

Net Cost Change 21-13+41-18+14-

16-13+41-27=17 25-13+41-18=35 18-14+18-27=-5 23-41+18-14=-14 32-18+14-17=11

Again, above table shows transportation cost can be reduced more by allocating some units into the cell

(O2, D4). Maximum 4 units can be shipped in that cell as shown in table 4. Origin\Destination O1 O2 17 6 O3 32 Demand Table 4 6 27 10 10 18 9 12 41 4 15 18 14 3 23 4 19 43 D1 21 D2 16 D3 25 D4 13 11 13 Supply 11

Next Iteration Of Stepping Stone Method. Unoccupied Closed Path Net Cost Change Remarks Cells (O1, D1) (O1, D1)-> (O1, D4)-> 21-13+23-17=14 Cost Increases (O3, D4)-> (O3, D1) (O1, D2) (O1, D2)-> (O1, D4)-> 16-13+23-14+1827=3 Cost Increases (O2, D4)-> (O2, D3)-> (O3, D3)-> (O3, D2)

(O1, D3) (O1, D3)-> (O1, D4)-> Cost Increases (O2, D4)-> (O3, D3) (O2, D2) (O2, D2)-> (O2, D3)-> Cost Decreases (O3, D3)-> (O3, D2) (O3, D1) (O3, D1)-> (O3, D3)-> Cost Increases (O2, D3)-> (O2, D1) (O3, D4) (O3, D1)-> (O3, D3)-> Cost Increases (O2, D3)-> (O2, D1)-> (O3, D1)-> (O3, D3)

25-13+23-14=21 18-14+18-27=-5 32-18+14-17=11 41-18+14-23=14

Again, above table shows transportation cost can be reduced more by allocating some units into the cell (O2, D2). Maximum 3 units can be shipped in that cell as shown in table 4. Origin\Destination O1 O2 17 6 18 3 14 23 4 D1 21 D2 16 D3 25 D4 13 11 13 Supply 11

O3 32 Demand Table 4 6 27 7 10 18 12 12 41 4 15

19 43

Next Iteration Of Stepping Stone Method. Unoccupied Closed Path Remarks Cells (O1, D1) (O1, D1)-> (O1, D4)-> Cost Increases (O3, D4)-> (O3, D1) (O1, D2) (O1, D2)-> (O1, D4)-> Cost Increases (O2, D4)-> (O2, D3) (O1, D3) (O1, D3)-> (O1, D4)-> 18=26 Cost Increases Net Cost Change 21-13+23-17=14 16-13+23-18=8 25-13+23-18+27-

(O2, D4)-> (O2, D2)-> (O3, D2)-> (O3, D3) (O2, D3) (O2, D3)-> (O3, D3)-> 14-18+27-18=5 Cost Increases (O3, D2)-> (O2, D2) (O3, D1) (O3, D1)-> (O3, D2)-> 32-27+18-17=6 Cost Increases (O2, D2)-> (O2, D1) (O3, D4) (O3, D4)-> (O3, D2)-> 41-23+18-27=9 Cost Increases (O2, D2)-> (O2, D4) Since all the unoccupied cells have positive values for the cost change, there is no way to improve the solution anymore. Hence optimum solution is as From Origin Transported Quantity Total Cost To warehouse F1 143 F2 102 F2 54 W1 W2 6 3 17 18 W4 11 Unit Cost

13

F2 92 F3 189 F3 216

W4 W2 W3

4 7 12

23 27 18 Rs.

Total Transportation Cost 796

Some special Cases 1. Multiple Optimum Solution: as with other linear programming problems, a transportation problem can have more than one optimum solution. This would be indicated when one or more of the unoccupied cells have zero value for the net cost change in the optimum solution. Thus the reallocation to the cell having a net cost change equal to zero will have no effect on the total transportation cost. This reallocation will provide another solution with the same total transportation cost, but the routes employed will be different from those from the original optimum solution. Multiple optimum solution are important because they provide management with added flexibility in management decision making.

Unbalanced Transportation problem: so far was been dealing with transportation problems in which the total supply at origins were equal to the total demand at destinations. In real life problems however, this may not be true. Thus the transportation problem with unequal demand and supply is called unbalanced Transportation Problem. Case 1: Supply Exceeds Demand:
m n

ai, > bj means total supply is greater than total demand.

i=1 j=1

In that case dummy destinations will be added to make balanced transportation problem. And the difference between the supply and the demand is considered as the demand for dummy destination. case 2: Demand Exceeds Supply
m n

ai, < bj demand.

i=1 j=1

means total supply is less than total

To make it balanced problem dummy row is to be added and excess demend is considered as the supply for that origin. Example: Consider following problem with initial basic solution:

Origin\Destination O1 8 O2

D1

D2 16 48

D3 16

Supply 152 164

32 1 O3 Demand Table 1

32 12 154

16 32 48 5 10 144 204 82

430

Solution: The above problem is unbalanced. Therefore introduce a dummy column with its requirement of 40 and cost of transportation in each cell is zero. the initial solution is obtained as in table 2 Origin\Destination D1 D2 D3 O1 8 16 16 70 82 O2 32 48 32 124 O3 Dummy Supply 152 0 164 0 40 154

Demand 470 Table 2 (it is left for students to verify that the solution given in the table 2 is optimum or not, if not then find the optimum solution).
2.

16 32 48 0 144 10 144 204 82 40

Degeneracy: it has already been established that an initial solution to an m origins and n destinations transportation problem consists of (m+n-1) allocations. But if the number of occupied cells is less than that at any stage of the solution, then the problem is said to have a degenerate solution. Degeneracy can occur at two stages: either (i) at the initial solution or (ii) during testing of the optimum solution. State 1: Degeneracy occur at initaial solution: to resolve degeneracy at the initial stage, we make use of an artificial quantity, denoted by in one or more of the unoccupied cells so that the number of occupied cells is m+n-1. The quantity is so small that it does not affect the supply and demand constraints. The value of is assumed as zero when it is considered in the movements of goods. Example: a manufacturing company has three

plants X, Y and Z, which supply to the distributors located at A, B, C. D and E. monthly plant capacities are 80, 50 and 90 units respectively. Monthly requirements of distributors are 40, 40, 50, 40 and 80 units respectively. Unit transportation cost are given below in rupees: From\ To X Y Z A 5 4 8 B 8 7 4 C 6 7 6 D 6 6 6 E 3 6 3

Determine an optimum distribution for the company in order to minimize the total transportation cost. Solution: Initial Solution by using Vogels Approximation method From\ To X Y 4 40 7 7 6 10 6 A 5 B 8 C 6 50 D 6 E 3 30 50 Capacity 80

Z 8 Dummy 0 Requirement 40
3.

90 4 40 0 40 6 0 6 3 50 30 0 0 30 50 40 80 250

Maximization Although the Transportation Problem which were dealt so far were of minimization, we may also come across cases of transportation problems with maximization objectives. When origins are related to destinations by a profit function instead of cost, the objective is to maximize case is to transform the profits to related cost. The transformation is conducted by subtracting the profit associated with each cell from the largest profit in the matrix. In this way the largest profit shows a zero relative cost, the objective changes to minimizing these relative costs. The solution procedures for relative costs are identical to those already described. After an OS to the transform minimization problem is determined, relative costs in occupied cells are restarted in their original profit functions, and the total profit is calculated according to quantities in the optimal distribution pattern.

Example: a company has 4 factories F1, F2, F3, & F4 manufacturing the same product production and raw material cost differ from Factory to factory and are given in the following table in the first two rows. The TC from the factories to the sales depots, S1, S2, S3 are also given. The last two columns in the table give the sales price and the total requirement at each depot. The production capacity for each factory is given in the last row. F1 F2 F3 F4 sales Requirement Per unit Production cost/unit15 Raw material cost/unit S1 3 Transportation S2 120 cost/unit S3 150 Supply 10 Determine distribution The deficit zero Solution: 18 10 9 1 5 14 9 5 7 8 13 12 9 4 34 80 4 5 32 3 6 31 price

150 50 100

the most profitable production and schedule and the corresponding profit. production should be taken to yield profit. The profit matrix is derived by the

equation: Profit = Sales price production cost raw material cost transportation cost Total requirement = 80 + 120 + 150 = 350 Total availabilities = 10 + 150 + 50 + 100 = 310 Dummy factory to produce 350 310 = 40 units. Table 1: (Profit matrix)

Factory/ Sales depot F1 F2 F3 F4 Requirement

S1 6 -2 3 8 80

S2 6 -2 2 5 120

S3 1 -4 2 3 150

Availability 10 150 50 100 310 350

Table 2: ( Balanced problem) Factory/ Sales depot F1 F2 S1 6 -2 S2 6 -2 S3 1 -4 Availability 10 150

F3 F4 FD Requirement

3 8 0 80

2 5 0 120

2 3 0 150 350

50 100 40 310

Table 3: (Relative Loss matrix) Factory/ Sales depot F1 F2 F3 F4 FD Requirement S1 2 10 5 0 8 80 S2 2 10 6 3 8 120 S3 7 12 6 5 8 150 Availability 10 150 50 100 40 350

Vogels method is now applied to derive the initial feasible solution as shown in the following Table 4 Table 4: Initial Solution VAM Factory/ Sales depot F1 F2 S1 2 10 10 10 12 150 S2 2 S3 7 Availability 10

40 F3 F4 80 FD Requirement 80 8 120 5 50 0 20 8 3 6

110 6 5 8 40 150 50 100 40 350

Since there are 7 (m+n-1) allocations, the initial solution is straightway tested for optimality in table 4 using MODI Method. Determine row and column numbers ui (I = 1, 2, 3, 4) and vj ( j = 1,2,3) by using the formula cij = ui + vj for occupied cells, as shown in Table 5. the opportunity costs for all the unoccupied cells are calculated by using formula ij = cij ( ui + vj ) and are also shown in the following table 5. Table 5: Initial Solution Non Optimal Factory/ Sales depot F1 F2 S1 2 3 10 10 3 (+) S2 2 3 12 110 150 u2 = 10 S3 Availability 7 10 Row Number ui u1 = 2

10 40 F3 2 F4 5 50 (-) 6

(-) 6 -2 (+) 50 u3 = 6

0 3 5 100 u4 = 3 80 20 0 FD 8 8 8 40 u5 = 0 5 2 40 Requirement 80 120 150 350 Column v1 v2 = v3 = Number vj = 0 2 -3 Since the cell (F3, S3) has the negative opportunity cost 2, it is admitted as an entering variable in the solution Result is shown in table 6

Table 6: Optimal Factory/ Sales depot F1 3 S1 2

Solution S2 2 10 3 S3 Availability 7 10 Row Number ui u1 = 4

10 3 F3 5 4 F4 0 80 FD 8 5 Requirement 80 Column v1 Number vj = -5

F2

10 90 6 2 3 20 8 2 120 v2 = 2

12 60 6 50 5 0 8 40 150 v3 = 0

150 50 100 40 350

u2 = 12 u3 = 6 u4 = 5 u5 = 8

The solution given above in Table 6 is an optimal one, since there is no negative opportunity cost in the unoccupied cells. The profit associated with this solution is : (80 x 8) + ( 10 x 6) + (90 x 2) + ( 20 x 5) + (60 x 4) + (50 x 2) = 480 4. Prohibited and Preferred Routes: In many practical situations, some of the routes in a transportation are prohibited for some external reasons such as unexpected floods, weight or size conditions, transportation strikes and road construction etc. such restrictions can be handled in the TP by assigning a very high cost( say M or

infinity) to the prohibited routes to ensure that these routes will not be included in the optimal solution and then the problem is solved in the usual manner. Example: A manufacturer must produce a product in sufficient quantity to meet contractual sales in next four months. The production capacity and unit cost of production vary from month to month. The product produced in one month may be held for sale in later months but at an estimated storage cost of Rs. 1 per unit per month. No storage cost is incurred for goods sold in the same month in which they are produced. There is no opening inventory and none is desired at the end of four months. The necessary details are given in the following table:

Month Cost Of

Contracted

Maximum

Unit

Sales Production Production 1 20 40 2 30 50 3 50 30 4 40 50 How much should the manufacturer produce month to minimize total cost?

14 16 15 17 each

Solution: The given problem can be converted in the following Transportation model: From/To 1 2 3 4 Requirements 1 14 M M M 20 2 15 16 M M 30 3 16 17 15 M 50 4 17 18 16 17 40 Availability 40 50 30 50

Since the total production exceeds the total sales, we introduce a dummy month with its requirements as (170 140) units having the transportation cost as zero. Then using VAM, an initial basic feasible solution is given in table 1. Since ther are 8 allocations at independent positions, we can proceed straightway to test this initial solution for optimality by finding numbers uI( I = 1, 2, 3, 4) and vj ( j = 1, 2, 3, 4, 5) such that cij = ui + vj for occupied cells. We also compute opportunity costs ij = cij (ui + vj ) for unoccupied cells. From/ To F1 S1 S2 2 2 S3 7 Availabilit Row y Number ui 10 u1 = 4

3 F2 3 F3 4 F4 8 0 FD

10 3 1 10 12 0 90 60 5 2 0 20 8 8 2 12 0 v2 =2 3 0 8 40 15 0 v3 =0 6 50 5 6

150 50 100 40 350

u2 = 12 u3 = 6 u4 = 5 u5 = 8

5 Requirement 8 0 Column v1 Number vj = -5

Assignment Problem The objective of assignment problem is to assign a number of origins(jobs) to the equal number of destinations(persons) at a minimum cost or maximum profit Mathematical Formulation
m n

Minimize (total cost) Z = cij xij

i=1 j=1

Subject to the constraints

n

xij = 1i, i=1,2,.,m (supply constraint)

j=1 m

xij = 1j, j =1,2,.,n (demand constraint)

i=1

and xij >= 0

Difference between Transportation Problem and Assignment Problem Transportation Problem No. of sources and no. of destinations need not be equal. Hence, the cost matrix is not necessarily a square matrix. xij, the quantity to be transported from ith origin and jth destination can take any possible positive values, and satisfies the requirements. The capacity and the requirement value is equal to ai and bj for the ith source and jth destinations (i = 1,2,..,m & j = 1,2,, n) Assignment problem Since assignment is done on one-one basis, the no. of sources and the no. of destinations are equal. Hence, the cost matrix must be a square matrix. xij, the ith job is assigned to the jth person and can take the value 1 or 0.

1.

2.

3.

The capacity and the requirement value is exactly one i.e., for each source of each destination the capacity and the requirement value is exactly one.

4.

The problem is unbalanced if the total supply and total demand are not equal.

The problem is unbalanced if the cost matrix is not a square matrix.

Hungarian method The solution of an assignment problem(minimization case) can be arrived using Hungarian ,method. The steps involved in this method are as follows: Step1. Find the opportunity cost table by: a. Subtracting the smallest number in each row of the original cost table or matrix from every no. in that row, and b. Then subtracting the smallest no. in each column of the table obtained in part (a) from every no. in that column. Step 2. Make assignments in the opportunity cost matrix in the following way: a. Examine the rows successively until a row with exactly one unmarked zero is found. Enclose this zero in a box() as an assignment will be made there and cross (X) all other zeros appearing in the corresponding column as they will not be considered for future assignment. Proceed in this way until all the rows have been examined.

After examining all the rows completely, examine the columns successively until a column with exactly one unmarked zero is found. Make an assignment to this single zero by putting square () around it and cross (X) out all other zeros appearing in the corresponding row as they will not be used to make any other assignment in that row, Proceed in this manner until all the columns have been examined. c. Repeat the operations (a) and (b) successively until one of the following situation arises: i. all the zeros in rows/columns are either marked ()) or (X) and there is exactly one assignment in each row and in each column. In such a case optimal assignment policy for the given problem is obtained. ii. There may be some row( or column) without assignment, i.e., the total number of marked zeros is less than the order of the matrix. In such a case, proceed to next step 4. Step 3: Revise the opportunity cost table Draw the minimum no. of vertical and horizontal lines necessary to cover all the zeros in the reduced cost table obtained from step 2 by adopting the following procedure:
b.

Mark ( ) all rows that do not have assignments. ii. Mark ( ) all columns ( not already marked) which have zeros in the marked rows. iii. Mark ( ) all rows ( not already marked) that have assignments in marked columns. iv. Repeat steps (ii) and (iii) until no more rows or columns can be marked. v. Draw straight lines through each unmarked row and each marked column. If the no. of lines drawn( or total assignments) is equal to the no. of rows(or columns) then the correct solution is the optimal solution, otherwise goto step 4.
i.

Step 4. Develop the new revised opportunity cost table: An optimal solution is seldom obtained from the initial opportunity cost table. Often, we will need to revise the table in order to shift one(or more) of the zero costs from its present location( covered by lines) to a new uncovered location in the table. Intuitively. We would want this uncovered location to emerge with a new zero opportunity cost. This is accomplished by subtracting the smallest no. not covered by a line from all numbers not covered by a

straight line. This same smallest no. is then added to every no. lying at the intersection of any two lines. Step 5. Repeat steps 2 to 4 until an optimal solution is obtained. Example 1: A compute center has got four expert programmers. The center needs four application programs to be developed. The head of the computer center, after studying carefully the programmes to be developed, estimates the computer time( in minutes) required by the respective experts to develop the application programmes as follows: Programmes A B C D 1 120 100 80 90 Programmers 2 80 90 110 70 3 110 140 120 100 4 90 90 80 90 Assign the programmers to the programmes in such a way that the total computer time is minimum. Solution: Step 1: Row Transformation: Subtract the smallest number from each row of the original cost table or matrix from every no. in that row

1 Programmers 3 4

Programmes A B C D 40 20 0 10 2 10 20 40 0 10 40 20 0 10 10 0 10 Table 1

Column Transformation: Subtracting the smallest number from each column of the original cost table or matrix from every no. in that row Programmes A B C D 1 30 10 0 10 Programmers 2 0 10 40 0 3 0 30 20 0 4 0 0 0 10 Table 2 Step 2: Starting with row 1 in table 2, examine rows one by one until a row containing exactly single zero element is found and make there assignment indicated by to that cell and simultaneously cross other zeros in that particular column. When the set of rows has been completely examined, an identical procedure is applied successively to

columns. Examine columns until a column containing exactly one remaining zero is found. Make an assignment in that position and cross other zeros in the row in which the assignment was made. Continue these successive operations on rows and columns until all zeros have been either assigned or crossed out. Thus, we get table 3 as shown below: Programmes A B C 1 Programmers 3 4 30 10 0 2 0 0 0 0 D 10

10 40 0 0 10

30 20 0

Table 3 It is clear from table 3 Hungarian method is not capable of doing all the allocations. So we select any zero at random and cross other zeros from respective row and column. Let us select zero at (2,1) position. Then again apply the same procedure on the table. As a result we get table 4. Programmes A B C D

1 Programmers 3 4

30 10 0 2 0 0 0 0

10

10 40 0 0 10

30 20 0

Table 4 Number of allocations are 4 same as the number of rows and number of columns. Hence the pattern of assignment among programmers and programs with their respective time (in minutes) is given below:

Programmer 1 2 3 4

Program C A D B

Time(in minutes) 80 80 100 90 Total time 350

Example 2: A company is producing a single product and is selling it through five agencies

situated in different cities. All of a sudden, there is a demand for the product in another five cities not having any agency of the company. The company is faced with the problem of deciding on how to assign the existing agencies to dispatch the product to needy cities in such a way that the traveling distance is minimized. The distances( in kms) between the surplus and deficit cities are given in the following distance matrix. Deficit Cities I II III IV V A 160 130 175 190 200 Surplus Cities B 135 120 130 160 175 C 140 110 155 170 185 D 50 50 80 80 110 E 55 35 70 80 105 Determine the optimal assignment schedule. Solution: Step 1: Row Transformation Deficit Cities I II III IV V A 30 0 45 60 70 Surplus Cities B 15 0 10 40 55 C 30 0 45 60 75 D 0 0 30 30 60 E 20 0 35 45 70

Table 1 Column Transformation Deficit Cities I II III IV V A 30 0 35 30 15 Surplus Cities B 15 0 0 10 0 C 30 0 35 30 20 D 0 0 20 0 5 E 20 0 25 15 15 Table 2

Step 2: Allocation Step A Surplus Cities C Deficit Cities I II III IV V 30 0 35 30 15 B 30 0 15 0 0 10 0 35 30 20

D E

20 0

20 0

25 15 15

Table 3 No. of allocations are less than the no. of rows or columns. Thus the optimal solution is not reached at this stage. Therefore, proceed to next step. Step 3: Draw the minimum no. of possible vertical or horizontal lines to cover all the zeros. i. Make the table 3 once again. ii. Mark () row 3 and 5 in which there is no assignment. iii. Then mark () column 2 which have zeros in marked rows 3 and 5. iv. Next mark () row 1 because this roe contains assignment in marked column 2. No further rows or columns will be required to mark during this procedure. v. Draw lines through marked columns and unmarked rows Deficit Cities I II III IV V

A C D E

30 0 30 0 0 0 20 0

35 30 15 0 10 0 35 30 20 20 0 5 25 15 15

Surplus CitiesB

15 0

Table 4 Step 4: Develop the revised Opportunity Cost table Select the smallest element (15), among all uncovered elements of the table 4, and subtract this value from all the elements of the matrix not covered by lines and add to every element that lie at the intersection of the lines and leaving the remaining elements unchanged. The following matrix is obtained. Again proceeding exactly in the same manner of step 3, we get Deficit Cities I II III IV V A 15 0 20 15 0 Surplus Cities C B 15 0 15 15 0 20 15 5 10 0

D E

0 5

15 20 0 0 10 0 Table 5

5 0

No. of allocations in table 5 are same as no. of rows, therefore the pattern of assignment among surplus cities and deficit cities with their respective distance( in kms) is given below: Surplus city A B C D E Deficit city V III II I IV Distance(kms) 200 130 110 50 80

Example 3: (Crew Problem): An airline that operates seven days a week has a time table shown below. Crews must have a minimum layover of 6 hrs between flights. Obtain the pairing of flights that minimizes layover time away from home. For any given pairing the crew will be based on the city that results in the smaller layover. Flight Delhi-Calcutta Calcutta-Delhi Flight

Depart Arrive No. Depart Arrive 1 7.00 am 9.00 am 101 9.00am 11.00am 2 9.00 am 11.00am 102 10.00am 12.00Noon 3 1.30 pm 3.30 pm 103 3.30 pm 5.30 pm 4 7.30 pm 9.30 pm 104 8.00 pm 10.00 pm For each pair also mention the town where the crew should be based. Solution: Layover times when crew based in Delhi: First of all construct the table for layover times between flights when crew is based in Delhi, we consider 30 minutes = I unit.

Flights 1 2 3

101 102 103 104

48 50 13 22 33 46 57 18 35 37 48 57

23 25 36 45

Table 1 Layover times when crew based in Calcutta Flights 1 2 3 4 101 102 103 104

40 38 27 18 44 42 31 22 53 51 40 31 17 15 52 43 Table 2

Now construct the table for minimum layover times between flights with the help of table 1 & 2. the layover times marked * denote that the crew is based on calcutta. Thus we get, table 3 as shown below: Flights 1 2 101 102 103 104

40* 38* 13 18* 44 42* 31* 18

3 4

35 37 40* 31* 17* 15* 36 43* Table 3

Row Transformation: Flights 1 2 3 4 101 102 103 104 5*

27* 25* 0 4 6

26 24* 13* 0 19* 0* 2* 0* 21 28* Table 4

Column Transformation:

Flights 1 2 3 4

101 102 103 104 5* 28*

25* 25* 0

24 24* 13* 0 2 6 19* 0*

0* 0* 21

Table 5 Reduced Table is: Flights 1 2 3 4 101 102 103 104 7*

25* 25* 0 0 4

22 22* 11* 0 17* 0* 30* 0* 0* 21

Table 6 No. of allocations in table 6 are equal to no. of rows. Therefore the pattern of optimal assignments among

routes with respect to the waiting time( in minutes) is given below:

Crew Based at Flight No. time(in mins) 1 Delhi 3-101 2 3 4 Calcutta 4-102 Delhi 1-103 Delhi 2-104

Waiting 390 540 1050 450

Maximization case Some assignment problems are phrased in terms of the maximizing the payoff profit, or effectiveness of an assignment instead of minimizing costs. It is easy to obtain an equivalent minimization problem by converting all numbers in the table to opportunity costs by subtracting every number in the original payoff table from the largest single number in that table. It turns out that the minimizing opportunity costs produces the same assignment as the original maximizing problem. Once the optimal assignment

for the transformed problem has been computed by the Hungarian method, the total payoff or profit is found by adding the original payoffs of those cells that are in the optimal assignment. Example: Five different machines can do any of the 5 required jobs, with different profits resulting from each assignment as shown in the adjoining table. Find out maximum profit possible through optimal assignment. Machine I II III IV V A 30 37 40 28 40 Jobs B 40 24 27 21 36 C 40 32 33 30 35 D 25 38 40 36 36 E 29 62 41 34 39 Table 1 Solution: Step 1: The given maximization problem can be converted into a minimization assignment problem by subtracting from the highest element as shown in table 2 Machine I II III IV V 32 25 22 34 22

Jobs

B C D E

22 22 37 33

38 30 24 0

35 29 22 21

41 32 26 28

26 27 26 23

Table 2

Row Transformation: Machine I II III IV A 10 3 0 12 Jobs B 0 16 13 19 C 0 8 7 10 D 15 2 0 4 E 33 0 21 28 Table 3 Column Transformation: Machine I II III IV A 10 3 0 8 Jobs B 0 16 13 15 C 0 8 7 6 D 15 2 0 0 E 33 0 21 24

V 0 4 5 4 23

V 0 4 5 4 23

Table 4 Allcations: I A Jobs B C D E 0 0 Machine II III IV V 0 8 0 16 13 15 4 8 7 0 6 0 5 4

10 3

15 2 33 0

21 24 23

Table 5

Revised Table: I A Jobs B C D 0 E 0 4 Machine II III IV V 0 2 0 8 1 4 0 12 9 3 0 11 0

14 3

19 2

37 0

21 24 23

Table 6 The following optimal assignment schedule is thus obtained. Job 1 2 3 4 5 Machine C E A D B Profit 40 36 40 36 62 Total 214

Unbalanced Problem The solution procedure to assignment problems just discussed requires that the number of rows in the

matrix equal the number of columns. Often, however the number of workers to be assigned does not equal to number of jobs listed in the columns. When this occurs and we have more roes than columns, we simply add a dummy column. If the number of jobs that need to be done exceeds the number of workers available, we add a dummy row. This creates a matrix of equal dimensions and allows us to solve the problem by Hungarian Method of assignment discussed above. Since the Dummy job is really nonexistent, it is reasonable to assign zeros element in its row or column as the cost estimate in the matrix. Example: Solve the adjoining unbalanced assignment problem of minimizing total time for doing all the jobs,

I 6

Job II III IV V 2 5 3 6

Operator

B C D E F

2 7 6 9 4

5 8 2 3 7

8 6 3 8 4

7 9 4 9 6

7 8 5 7 8

Determine the optimal assignment schedule. Solution: Since the number of operators are not equal to the number of jobs, a dummy job 6 is created. The time consumed by any operator for the dummy job is zero. Job I II III IV V VI A 6 2 5 3 6 0 Operator B 2 5 8 7 7 0 C 7 8 6 9 8 0 D 6 2 3 4 5 0 E 9 3 8 9 7 0 F 4 7 4 6 8 0 Column Transformation Job I II III IV V A 4 0 3 0 1 Operator B C 0 5 3 6 5 3 5 7 2 3

VI 0 0 0

D E F

4 7 2

0 1 5

0 5 1

2 7 4

0 2 3

0 0 0

Number of allocations is less than the number of rows. Hence next table is

A Operator B C D E

I 4 0 4 4 6

Job II III IV V 0 2 0 1 3 5 0 0 5 2 0 4 5 6 2 6 2 2 0 1

VI 1 1 0 1 0

Number of allocations is equal to the number of rows. Hence Solution is Operator 1 to Job 4, Operator 2 to Job 1, Operator 3 to Dummy 6, Operator 4 to Job 5, Operator 5 to Job 2, Operator 6 to Job 3. Restricted Problem It is sometimes possible that a particular person is incapable of doing certain work or a specific job cannot be performed on a particular machine, the solution of the assignment problem should take into account these restrictions so that the infeasible assignments can be avoided. This can be achieved by assigning a very high cost (written as M or infinity) to the cells where assignments are prohibited. This also prohibits the entry of this pair of resourceactivity into the final solution. Example: In the modification of a plant layout of a factory four new machines M1 , M2, M3,and M4 are to be installed in a machine shop. There are five vacant places A, B, C, D and E available. Because of limited

space, machine M2 cannot be placed at C and M cannot be placed at A. the cost of locating of machine I to place j in rupees is shown below. Find optimal assignment schedule. A M1 M2 M3 M4 B 9 12 14 C 11 9 11 8 D 15 14 12 E 10 10 11 7 11 9 7 8

Solution: Steps 1. As there five vacant places and four machines, we take one dummy machine MD with zero cost of installation. Further the cost for cells M3A and M2C which are not to be filled is taken as M. Accordingly the cost matrix will be : From/ToA M1 M2 M3 M4 MD B 9 C D E

11 15 10 11 M 10 9 12 7 0 0 8 0

12 9 14 8 0 0

M 11 14 11 7

Step 2: Subtracting the minimum element from each row and each column and allocating assignments, we obtain. From/ToA M1 M2 M3 M4 MD B 0 3 7 0 C 2 0 1 0 D 6 7 5 0 E 1 4 0 0 2 0 0 1 0

M 1

M 4

Since each row and column is having only one assignment, the optimal assignment is reached and is given by: M1 A M2 B M3 E M4 D (Space C will be kept vacant) with minimum cost = 9 + 9 + 7 + 7 = 32

Alternate Solutions Sometimes, it is possible to have two or more ways to cross out all zero elements in the final reduced matrix for a given problem. This implies that there is more than the required number of independent zero elements. In such cases, there will be multiple optimal solutions with the same total cost of assignment. In such type of situations, management may exercise their judgment or preference and select that set of optimal assignments, which is more, suited to their requirement. Example: A company has four zones open and four salesman available for assignment. The zones are not equally rich in their sales potential. It is estimated that a typical salesman operating in each zone would bring the following weekly sales: Zone : I II III

IV Weekly Sales(Rs.) : 63,000

1,26,0001,05,00084,000

The four salesmen are also considered to differ in ability. It is estimated working under the same condition, their weekly sales would be proportionately as follows:

Salesman : A B C D Proportion : 7 5 5 4 If the criterion is maximum expected total sales, the intuitive answer is to assign the best salesman to the richest territory, the next best salesman to the second richest and so on. Verify this answer by the assignment technique. Solution: Step 1: Take Rs. 1000 as one unit. Total weekly sales are 21. Divide the individual sales in each territory by 21 to obtain the required annual sales by each salesman. The maximum sales matrix is thus obtained as Zone 101 102 103 104 1 Salesman 3 4 42 35 28 21 2 30 25 20 15 30 25 20 15 24 20 16 12 Table 1 Step 2: Converting maximization problem into minimization problem

Zone 101 102 103 104 1 Salesman 3 4 0 2 7 14 21 12 17 22 27

12 17 22 27 18 22 26 30

Table 2 Step 3: Apply Row and Column Transformation, we get

Zone 101 102 103 104 1 0 3 0 1 0 6 1 2 0 9 2 3 0 3

Salesman 2 3 0 4 0

 Table 3 Zone 101 102 103 104 1 0 2 0 0 0 5 0 1 0 8 1 2 0 2

Salesman 2 3 0 4 1

Table 4 Zone 101 102 103 104 1 0 2 0 0 1 4 0 0 0 7 0 1 0 1 Salesman 2 3 0 4 1

Table 5 There are two zeros in 2nd & 3rd rows and 102nd & 103rd columns. Therefore assignments can be done in two different ways as shown in table 6 & table 7.

Zone 101 102 103 104 1 0 2 0 0 1 4 0 0 0 7 0 1 0 1 Salesman 2 3 0 4 1

Table 6

Zone 101 102 103 104 1 0 2 0 0 1 4 0 0 0 7 0 1 0 1 Salesman 2 3 0 4 1

Table 7 Assignment Set I Assignment II Salesman Zone Sales (000s) Salesman Zone Sales (000s) A I 42 A I 42 B II 25 B III 20 C III 20 C II 25 D IV 12 D IV 12 Total 99 Total 99

Example: A firm produces four products. There are four operators who are capable of producing any of these four products. The processing time varies from operator to operator. The firm records 8 hours a day and allows 30 minutes for lunch.

Traveling Salesman Problem The traveling salesman problem can be stated as follows: Find the lowest cost route for a traveling salesman (if travel expense per mile are constant, then the lowestcost route is the shortest) who must leave a given city stop at every city in a certain region and then return to his starting point. This type of problem arises in the following management phenomenon. Deliveries Inspection School Bus Routing Television Relays Let us outline a very simple case: what is the lowestcost regular route for a traveling salesman who has to

stop at cities A, B, C and D without passing twice through the same one B ABCDA = 12 + 17 + 30 + 23 = 82 12 14 A 25 C 23 D In tabular form problem can be described as A B C D A 0 12 25 23 B 12 0 17 24 C 25 17 0 30 D 23 14 30 0 No analytical method is known for solving problem like this and it is impossible in practice to calculate all the solutions in order to choose the best when we are dealing with a large number of cities. For example with 12 cities there are 19958400 routes are possible. One can also imagine a case in which routes must be planned within an incomplete network i.e., where there are no connections between certain cities. 30 17
ABDCA = 12 + 14 + 30 + 25 = 81 ADBCA = 23 + 14 + 17 + 25 = 79 ACBDA = 25 + 17 + 14 + 23 = 79 Thus the route ADBCA gives the lowest cost.

In this case the problem becomes still more complicated. Example: A salesman has to visit five cities A, B, C, D and E. the intercity distances are tabulated below: From/ToA B C D E A - 12 24 25 15 B 6 - 16 18 7 C 10 11 - 18 12 D 14 17 22 - 16 E 12 13 23 25 Note that the distance( in kms.) between two cities need not be same both ways. Required: if the salesman starts from city A and has to come back to city A, which route would you advise him to take so that total distance traveled by him is minimized? Solution: Step 1: Row Transformation From/ToA A B C D E B C D E

M 0 0 0 0 0 1 3 1

12 13 3 M 8 8 2

M 10 12 1 M 2

11 13 M

Step 2 : Column Transformation From/ToA A B C D E B C 4 D 5 4 E 2 0 1 M

M 0 0 0 0 0 1 3 1

M 2 0 3

M 0 5

M 1

From/ToA A B C D E

Step 3: Allocation C D E 4 5 4 2 0 1 M

M 0 0 0 0 0 1 3 1

M 2 0 3

M 0 5

M 1

Now since each row and each column has one enrectangled assignment, an optimal solution is reached. The optimal assignments are: A to B + B to E + C to D to C + E to A = 12 + 7 + 18 + 22 + 12 = 71 km. Based on above, total distance traveled will be minimum. This assignment schedule does not provide the solutions of traveling salesmen problem as it gives A to B, B to E and E to A, while E is not allowed to follow A, till C and D are processed. Therefore, we try the next best solution, which satisfy this extra (traveling salesmen problem). Now, the next minimum (non-zero) element in the matriux is 1. so we try into the solution. We start with making an assignment at (5,4) instead of zero assignment at (4,3). The resulting solution is shown in following table From/ToA A B C D B C 4 D 5 4 E 2 0 1

M 0 0 0 0 1 3

M 2 0

M 0

M 1

The Optimum Schedule is: A B, B C, C D, D E, E A involving distance of 12 + 16 + 18 + 16 + 12 = 74 km. This solution satisfies the traveling salesman problem. There can be another schedule also, i.e., A B, B E, E C, C D, D A involving distance of 12 + 7 + 23 + 18 + 14 = 74 km. This route also satisfies the traveling salesman problem.