ASSIGNMENT PROBLEM
ASSIGNMENT PROBLEM
If m n and ai bi 1 for i 1, n. The Assignment problem is : n n Minimum x c .x o ij ij i 1 j 1 n Subject to x 1; i, j 1, n; ij j 1 n x ij 1; i, j 1, n; i 1 x 0 or 1. ij
The Hungarian Method Provides a simple heuristic that can be used to find the optimal set of assignments. Is easy to use, even for fairly large problems. Is based on minimization of opportunity costs that would result from potential pairings.
These additional costs would be incurred if the lowest-cost assignment is not made, in terms of either jobs (i.e., rows) or employees (i.e., columns).
Special Situations Special Situations
Certain situations can arise in which the model deviates slightly from that previously described.
Among those situations are the following:
The number of rows does not equal the number of columns. The problem involves maximization rather than minimization. Certain matches are undesirable or not allowed. Multiple optimal solutions exist.
Assignment problem
We will explain the Hungarian Algorithm of Assignment Problem by solving an example. Example : A department head has four subordinates and four tasks to be performed. The subordinates differ in efficiency and the tasks differ in their intrinsic difficulty. His estimate of time each man would take to perform each tasks is given in the effectiveness matrix below. How would the tasks be allocated, one to a man, so as to minimize the total man hours.
Assignment Problem
Tasks Subordinates
I 8 13 38 19
II 26 28 19 26
III 17 4 18 24
IV 11 26 15 10
A B C D
Assignment
Step 1: Subtract the smallest element in each row from every element in that row. We get the reduced matrix Step 2: Next subtract the smallest element in each column from every element in that column. We get the second reduced matrix (Steps 1 & 2 will bring at least one zero in every row and column) Step 3: Test whether it is possible to make an assignment. If it is possible then, the assignment is optimal. For testing: (A) Examine first row. If there is only one zero, then surround this zero by a square [0] and cross (X) all zeros in the column passing through this surrounded zero. Next examine second row, third row etc. and repeat above for every row having one zero. If there are more than one zeroes in any row, then do not touch this row and pass to next row. (B) Now repeat the above procedure for columns (i.e. in (A) replace row by column and column by row). We continue these successive operations [3 (A) and 3 (B)] on rows and columns until all zeros are either crossed or surrounded.
Assignment Problem
A
MS. P MS.Q MR.R MS.S MR.T
B 130 120 110 50 35
C 175 130 155 80 70
D 190 160 170 80 80
E 200 175 185 110 105
160 135 140 50 55
Assignment
Step 4: Draw (by trial) minimum number of vertical and or horizontal lines to cover all zeros. (Let n denote the number of rows in the problem. By theory if m<n then allocation at this stage is not possible, if m=n it is possible). Proceed as follows: (A) Select the smallest element, say x, among all elements uncovered by lines. (B) Subtract this value x from all elements not covered by lines and x to all those elements that lie at the intersection of lines.
TRANSPORTATION PROBLEM
A TRANSPORTATION PROBLEM INVOLVES TRANSPORTING ITEMS FROM SOURCES WITH FIXED SUPPLY TO DESTINATIONS WITH FIXED DEMAND AT LOWEST COST.
TRANSPORTATION PROBLEM
A PRODUCT IS TO BE TRANSPORTED FROM m SOURCES Si TO n DESTINATIONS Dj ai = QUANTITY AVAILABLE AT Si bj = QUANTITY REQUIRED AT Dj cij = COST OF TRANSPORTATION OF ONE
UNIT FROM SOURCE Si TO DESTINATION Dj
HOW MUCH SHOULD BE TRANSPORTED FROM EACH SOURCE TO EACH DESTINATION SO AS TO MINIMIZE COSTS.
TRANSPORTATION PROBLEM
Minimum x o cij . xij
i 1 j 1 n
.... (1)
Subject to x ij ai , i 1, m; ai 0....(2)
j1 m
x
i 1 m
ij
b j , j 1, n; b j 0.... (3) ............ (4)
j
x ij 0 i, j
ai . b
i 1 j 1
................ (5)
' m n - 1' independen t constraint s.
TRANSPORTATION PROBLEM
D1
S1 S2 S3 5
D2
2
D3
4
D4
3 60
6
2 50 65
4
3 65
9
8 30
5 60
1 90
T RANSPORTATION PROBLEM
1) NORTH WEST CORNER RULE (NWCR) TC=RS.1045 2) LEAST COST METHOD (LCM) TC=RS. 845 3) VOGELS APPROXIMATION METHOD (VAM) TC=RS. 665 m+n-1 NON-ZERO ENTRIES (INDEPENDENT CONSTRAINTS) BALANCED / UNBALANCED PROBLEMS TEST FOR OPTIMALITY (Ui +Vj - Cij 0)
TRANSPORTATION PROBLEM
National Generators
ORIGIN PLANT 3-MONTH PRODUCTION CAPACITY (UNITS) 5000
MUMBAI
NEW DELHI
6000
KOLKATA
2500 -------TOTAL 13,500
TRANSPORTATION PROBLEM
DESTINATION
DISTRIBUTION CENTER CHENNAI
3-MONTH DEMAND FORECAST (UNITS) 6000
NAGPUR
4000
KANPUR
2000
BHOPAL
1500 -------TOTAL 13,500
TRANSPORTATION COST PER UNIT FOR THE NATIONAL GENERATORS TRANSPORTATION PROBLEM
DESTINATION ORIGIN MUMBAI NEW DELHI KOLKATA CHENNAI 3 7 2 NAGPUR 2 5 5 KANPUR 7 2 4 BHOPAL 6 3 5
NETWORK REPRESENTATION OF THE NATIONAL GENERATORS TRANSPORTATION PROBLEM
DISTRIBUTION CENTERS (DESTINATION NODES) PLANTS (ORIGIN NODES) TRANSPORTATION COST PER UNIT
1 CHENNAI
6000
5000
1 MUMBAI
2 NAGPUR
4000
6000
2 NEW DELHI 3 KANPUR
2000
2500
3 KOLKATA
4 BHOPAL
1500
SUPPLIES
DISTRIBUTION ROUTES (ARCS)
DEMANDS
TRANSPORTATION PROBLEM
Min s.t .
3x11 2x12 7x13 6x14 7x 21 5x 22 2x 23 3x 24 2x 31 5x 32 4x 33 5x 34 x 11 x 12 x 13 x 14 x 21 x 22 x 23 x 24 x 21 x 12 x 13 x 14 x ij 0 for i 1, 2, 3; j 1, 2, 3, 4 x 22 x 23 x 24 x 31 x 32 x 33 5000 6000 x 31 x 32 x 33 x 34 2500 x 11 6000 4000 2000 x 34 1500
OPTIMAL SOLUTION TO THE NATIONAL GENERATORS TRANSPORTATION PROBLEM
ROUTE FROM TO UNITS SHIPPED 3500 1500 2500 2000 1500 2500 COST PER UNIT (RS IN 1000) 3 2 5 2 3 2 TOTAL COST (RS. IN 1000) 10,500 3.000 12,500 4,000 4,500 5,000 --------39,500
MUMBAI MUMBAI NEW DELHI NEW DELHI NEW DELHI KOLKATA
CHENNAI NAGPUR NAGPUR KANPUR BHOPAL MUMBAI
TRANSSHIPMENT PROBLEM
Transportation With Transshipment:
Example: Solve the following with Transshipment
S1 S2
D1 2 4
200
D2 5 2
150
D3 3 6
150
200 300
Where the Cost Matrices are: S1 S2 D1 D2 S1 0 2 D1 0 1 S2 2 0 D2 1 0 D3 2 4
D3 2 4 0
The problem is reduced to the following Transportation Problem:
S1 S2 D3 D4 D5 S1 0 2 2 5 3 500 S2 2 0 4 2 6 500 D3 2 4 0 1 2 700 D4 5 2 1 0 4 650 D5 3 6 2 4 0 650 700 800 500 500 500
TRANSPORTATION COSTS PER UNIT FOR THE EELECTRONICS TRANSSHIPMENT PROBLEM
WAREHOUSE PLANT MUMBAI NEW DELHI NAGPUR 2 3 KANPUR 3 1 RETAIL OUTLET WAREHOUSE NAGPUR KANPUR BHOPAL 2 4 KOLKATA 6 4 JAMSHEDPUR 3 6 BANGALORE 6 5
NETWORK REPRESENTATION OF THE EELECTRONICS TRANSSHIPMENT PROBLEM
PLANTS (ORIGIN NODES) WAREHOUSES (TRANSSHIPMENT NODES)
RETAIL OUTLETS (DESTINATION NODES)
5 BHOPAL
2OO
6OO
1 MUMBAI
3 NAGPUR
6 KOLKATA
15O
7 JAMSHEDPUR
4OO
35O
2 NEW DELHI
4 KANPUR 8 BANGALORE
DISTRIBUTION ROUTES (ARCS)
3OO
SUPPLIES
DEMANDS
LINEAR PROGRAMMING FORMULATION OF THE E-ELECTRONICS
Min 2x13 + 3x14 + 3x23 + 1x24 + 2x35 + 6x36 + 3x37 + 6x38 + 4x45 + 4x46 + 6x47 + 5x48 s.t. x13 + x14 x23 + x24 x13 x14 x23 + x35 + x36 + x37 + x38 x24 x35 x36 + x45 + x46 + x47 + x48 + x45 + x46 600 400 =0 =0 =200 =150
Destination node constraints origin node constraints
Transshipment node constraints
x37
x38 xij 0 for all i and j
+ x47
+ x48
=350
=300
OPTIMAL SOLUTION TO THE E-ELECTRONICS TRANSSHIPMENT PROBLEM
ROUTE
FROM MUMBAI MUMBAI NEW DELHI NAGPUR NAGPUR KANPUR KANPUR
TO NAGPUR KANPUR KANPUR BHOPAL JAMSHEDPUR KOLKATA BANGALORE
UNITS SHIPPED 550 50 400 200 350 150 300
COST PER UNIT IN THOUSANDS 2 3 1 2 3 4 5
TOTAL COST IN THOUSANDS 1100 150 400 400 1050 600 1500 ----------5200
PROBLEM VARIATIONS
TOTAL SUPPLY NOT EQUAL TO TOTAL DEMAND MAXIMISATION OF OBJECTIVE FUNCTION ROUTE CAPACITIES OR ROUTE MINIMUMS UNACCEPTABLE ROUTES.
A GENERAL LINEAR PROGRAMMING MODEL OF THE TRANSSHIPMENT PROBLEM
The general linear programming model of the transshipment problem is Mi n c x ij ij all arcs s.t. x x s Origin nodes i ij ij i arcs out arcs in x x ij ij 0 arcs out arcs in Transshipment nodes
x d Destination nodes j x ij ij j arcs in arcs out x ij 0 for all i and j
Where xij = number of units shipped from the node i to node j cij = cost per unit of shipping from node i to node j si = supply at origin node i dj = demand at destination node j
