chapter 3 Assignment Model Assignment Model -Introduction «Hungarian method for Assignment problem m introduction to Assignment Problems ~ pe asinment problem is an allocation problem. It is a special type of transportation problem in which the per of sources (Jobs or factories) equals to the number of destinations (Machines or warehouses), and all jemand and capacity corresponding to destinations and sources equals to one unit. assignment problems are useful in, {) Decision theory i) Production control i) Defence operation ix) Manpower management. . The objective of assignment problem is to assign a number of origins (jobs) to the number of destinations (persons or machines) at a minimum cost (for maximum profit). + Toexamine the nature of assignment problem, suppose there are n jobs to be performed and n persons are tailable for doing these jobs. Assume that each person can do each job at a time, though with varying degree of efficiency. Let Cy be the ‘est (payment) if the i person is assigned to j job. The problem is to find an assignment (which job should be assigned to which person) so that the total cost for Yerforming all jobs is minimum. Problems of this kind are known as assignment problems. The assignment problem can be stated in the form nx n matrix [Cy], which is called as cost-matrix where the entry Cy denote the cost when i person is “ssigned to jt job.Consider the cost matrix [Cy] where, Ci is Table 3.2-1 ne 2 [Cx Ca Person | 3 [Cn Ca n {Cn Ca = 3.2__ Mathematical Representation of Assi nment Problem ~y the cost when i person is assigned to j% job as follow, a ee Suppose xy represents an assignment when i* person is assigned to perform j? Job. Hence —xy_= 0, if it* person is not assigned to j® job. = 1 if i® person is assigned to j* job. ‘Thus we have to, Minimize the total cost i.c. +n) = 1 (one job is done by the i* person, i= 1, 2, »n) 1 (only one person should be assigned to j* job, j = 1, 2, 'M & ; and xy 2 O (ie. xy =0 or xy= 1). 3.3. Hungarian Method of Solving A.P. (Minimization Case) 3.3.1 Basic Assumption () Each facility is capable of performing each job. (ii) Facilities differ in their capabilities and jobs differ in their work contents. (iii) Only one job can be assigned to each facility. (iv) Assignment matrix must be a square matrix (If not dummy row or column are added to make * sous matrix). __fi 1 . jon of the assignment prob) P . A p | solution the en be ems remains the same if a constant is added or subtracted to any fi the cost MALTS 5 Le. assignment problem if a constant is added or subtracted to every sn mn) of the cost matri i al vow cor eal ) of matrix then an assignment which minimizes the total cost for one matrix, pa tne total cost for the other matrix. site 2 qzoand there exists a solution xy = Xj wv non nat his solution is an optimal solution, steps to Solve Assignment Problem by Hungarian Method 11 Consider the given cost matrix. Check whether it contains same number of rows and columns #7 tmerwise add dummy row or column with zero cost. a+ Select smallest element in each row and subtract i from all elements of that row. yp3:_ Select smallest element in each column of reduced matrix obtained from step 2 and subtract it from all clements of that column. pt: Make the assignment for the reduced matrix obtained from steps 1 and 2 as follows: 2) Examine the rows successively until a row with exactly one zero is found. Make an assignment to this ‘zero’ by putting square () around it and strike off (X) all other zeros appearing in the corresponding column, as they will not be used to make any other assignment in that column. Continue the process until all rows have been examined. b) Examine the columns successively until a column with exactly one zero is found. Make an assignment to this zero by putting square |) around it and strike off (X) all other zeros appearing in the corresponding rows as they will not be used to make any other assignment in that row. Continue this until all the columns have been examined. ©) Repeat steps 4(a) and 4(b) until all zeros in rows /columns are either marked U or strike off (X). If the number of assignments ( marked ()) made are equal to the order of the cost matrix, then it is an optimal solution, otherwise go to step (5). Draw the minimum number of horizontal and vertical lines to cover zeros in the reduced matrix obtained form step (4) as follows : ®) Mark (¥) all rows for which the assignment has not been made. ») Mark (v) all columns that have zeros in marked rows. ©) Mark (¥) all rows that have assignments in marked columns. tos: ® Repeat steps 5(a) of 5(c) until no more rows or columns are left. S20 straight lines through all unmarked rows and marked columns.ler of the cost matrix, then it ig gy ibtract this smallest elemen, s ered elements. Sul t Step 7: Select the smallest element among all the uncov' ioc ot Te from ay the uncovered elements and add it tothe element which lies at the inte obtain another reduced matrix for fresh assignments. ‘Step 8: Goto step (4) and repeat the procedure until an optimal solution Step 9: Write down the optimal assignment corresponding to assigned 7er0s- Step 10: Find optimal cost, if needed by adding the costs at assigned points in the cos We illustrate this method by following examples. If the number of straight lines drawn are equal to ord solution, otherwise goto step (7). Step Hence ig is obtained. t matrix. Ex. 3.3.1 : Solve following assignment problem for minimization : abcd Af1l4 6 3 B|9 7 10 9 c]45 17 ols7s85 Soin. : In the given A.P. Number of rows = Number of column = 4 Step: i.e. given cost matrix, which is a square matrix. To find optimum solution using Hungarian method. bed 463 710 9 7 5 785 voup DAO p Step 2: Select smallest element in each row and subtract it from all elements of that row. This step gives anew assignment matrix as follows abed AO 3 5 2 Bi/203 2 cj/0173 DL3 23 0) Step 3: Now select smallest element in each column from above matrix and subtract it from all elements of that column. This gives new matrix as follows. abed AO 322 B/200 2 cClo143 DLs 20 oJnex v elect ~ow which has only one zero, Here 1 h Tow is a row whic i le Zero, 80 we which contains only one r t , 80 we make ) (Q) to this zero and cross other zero in the corresponding column. (1* column) ence we Rave: vouwDr wrHowt on ON? owen ne ox Ele ow tere is no row which has only one zero, ) c— — pa has only one zero. We find that 2 column is a column which contain fa ro. So we make an assignment to this i i ears zero and cross other zeros in the corresponding row sep 4e) #ARter this we find that 3" column is a column whieh contains only one zero (since one zero in Ps . - rewi and 3* column is crossed in previous step) We make an assignment to this zero and cross other zeros in the corresponding row (4* row). 204 row . We have assignment as follows a bec a A (pol Gee 2 Bal eene [Ola cl wi 4 3 DLs 2 fo «x Here number of assignment = 3 which is not equal to order of matrix. .. This is not an optimal solution Step 5(a) : Now mark the row which do not have assignment c. we mark 3* row. Step 5(b) :In this row, crossed zero occurs at 1* column. we mark 1* column. Step 5(c) :In the 1* column, assignment o°curs at 1* row, we mark 1* row- ‘we have marking as follows : fapabyetcnnd| va) D3 2 2 B} 2 O «x1 vc] m1 4 3 D 320KOperation Research (SPPU. (Comp. 2) sed. No more row / column ten, ee Step 5(€) :All zeros from rows/ columns are either assigned ha {4% rows) and marked column (14 Step 5(e) : Draw the straight lines through unmarked rows ( oh ¥ that all zeros are covered (assigned / erossed). ‘ =. We draw the straight lines as follows : + his is Step 6: Here the number oflines = 3 which is not equal to order of matrix, hence this is not optimal soy, Step 7: Select the smallest element among all uncovered elements i. select an element 1 from clements and subtract it from all uncovered elements, add it at point of intersection of tw jn, remaining elements keep as it. This step gives new assignment matrix for fresh assignment as follows : ad abcd Aro 211 B\3 00 2 clo 0 3 2 pl4200 Step 8: We make assignment as follows : Here 1* row is a row which contains only one zero, hence we make an assignment to this zero and eng other zeros in the corresponding column (1* column) ‘After this we find 3" is a row which contains only one zero (since zero in the first column is cross We make assignment to this zero and cross other zero in the corresponding column (2 column) ‘Now, we find that 2! row is a row which has only one zero, (“other zero in 2%¢ column is crossed): make an assignment to this zero and cross other zero in the corresponding column (i.e. cross zero in 3r4 column) At last we find that 4" row a row which contains only one zero (““ other zero in 3° column is eros We make an assignment for this zero. Hence we have assignments as follows : a booed aj @ 2 11 B} 3 w& [Oo] 2 Clea (olpeseea D 4 2 «x f Here number of assignments equals to order of matrix. +» This is an optimal solution, Hence we have assignments as Asa, BoeC>b,D34 and optimal cost = 1+ 10+5 +5=21 Rs.Ex. 3. ines Mi, Mz, Ms and Ma. Th i 3.4 Jobs J, J, Js, Js are to be assigned to four machines My, Mz, Ms and . le Processing ¢ Fi i all proces: | given below. Find allocation that minimizes the overall processing co* Soin. : Consider the given cost matrix. My jd [10 h} 5 J3 12 48 M uf 10 hh | 5 Js | 12 Je 8 Here number of machines = number of jobs = 4 +. The given A. P. is balance of order 4 M, 2 10 14 1s M2 12 10 4 ay To find optimum solution using Hungarian method. St (in Ry), i Mg ll 8 1 9 M3 19 7 13 n Ms Ma 19 1 13 i i Step 1: Subtracting row minimum from respective row elements we get reduced matrix as, Mi MMs Mc APO 29 1 hi0 5 23 Js }1 3 20 wLO73 1 Step 2: Subtracting column minimum from respective column elements. ‘Mi M2Ms Mi hpoo7i h}0 3 0 3 &b)/1 100 &LO5 11 Step 3: We make assignment as follows : M: Mz Ms My xf XM 7 1 as] K 3 OD 3 Boi 1 «x J ‘LOM 5 11 Hence number of assignments = order of the matrix = 4 ‘The problem has optimal solution given by J1—> Mo, Jz» Ms, Js—> Ma, Je—> Mi and minimum cost 12+74+11+8 38 Rs.eave the following assignment problem for minimization. a M: M, M3 My 412 30 21 15 J | 18 33 09 31 Js | 44 25 24 21 Je L23 30 28 14 » sposiven AP. Number ofrows = Number of column Db =4 ven A. P- is balance of order 4 qofind optimum solution using Hungarian method. 0 1: subtracting smallest element from each row respectively gpl! M: Mz Ms Ma ho w 9 3 Jo] 9 24 0 22 Js |23 4 3 0 ilo 1 14 0 ‘ep 2? Subtracting smallest element from each column respectively : M: M2 Ms Mc ho uw 9 8 | 9 2 0 2 i] 23 0 3 0 Je 9 12 14 «0 itep3: We make an assignment as follows : M: Mz Ms Mi o} 144 9 3 Bn melt 29) Oe 110) 22 Js] 23 fl 3 K so 9 2 u Hence number of assignments = order of the matrix. + This is an optimal solution + Optimal assignment schedule is J. Mada Mads > Ma, Ju Me = total cost = 12 +09 +25 +14 = 60 units.3.4 Maximization in Assignment Problems + In some problems we want the assignment such that the total optimum value is maximum, for example problems with outputs, profit, sales ete. + Hence we have an assignment problem in which the objective function is to maximize the total payoff. + To solve the maximization problem, firstly we convert the given maximization problems into minimization problem and this minimization problem we solve by using Hungarian method. Steps to solve maximization problems are as follows : i) Select the largest element of the given cost matrix. ii) Subtract each element of the given matrix from this largest element. iii) The resulting matrix is the new matrix for minimization assignment problem. iv) Solve this assignment problem for minimization by Hungarian method. * Another method of converting maximization problem to minimization problem is to multiply each element the cost matrix by — 1, which gives the matrix for minimization assignment problem.___ Districts 1alelelole] 1 | 32 | 38 | 40 | 28 | a0 2 | 40 | 24 | 28 | 21 | 36 33 | 30 | 37 fend | 41 | 36 | 36 5 | 29 [33 | 40 [35 39 | find an optimal assignment to maximize the profit. ida eee 7 igegiven AP. Number of salesman = Number of districts =5, 7 qhegiven A. P. is balance of order 5. sitisa maximization problem we locate the maximum entry and subtract each entry from it and solve it for jiization. A|slclple 1,9|3 1)13;1 a, iia) 1G 3/0/14] 8 iia 4ji9[s[olsls [13 8 1\/6\/2 ‘Tofind optimum solution using Hungarian method. #1: Subtract smallest entry from each row from respective rows a|B|c|ple 1jsj2{o|2lo 2| 0 |16| 12/19] 4 3|o|14|s8|ula 4|i9] 3 [o|5|5 5)12)7/0)5)1 "2: Now subtracting smallest entry from each column from respective columns, Alp lc[pl[E a een Eoy non e7a|c0 2) 0 | 14} 12) 14) 4 3) 0)12)8)6\4 4{19| 1/0 0 5 NL 5|12| 5 | 0 1Step 3: We make assignment as follows 2|(O] | 14 | 12 | 4 | 4 wii) 3s | 6 | 4 4}i9} 1 | x [|S 5/az| 5 |f | Kj 1 As number of assignments are not equal to order of matrix. + This is not an optimal solution Step4: We mark rows and columns as follows : w|i] 14) 4 5 ste. 12 | 8 ie mK Het ‘Here number of lines are not equal to order of matrix. ‘This is not optimal solution. Ble Step 5: Subtracting smallest uncovered element ie. 4 from all uncovered elements and add it to intersection we get fresh cost matrix as follows, A B c D E 1} 2\/b) |x) 7\x 2) (0) 10 8 lo | WK 3)/K)sla)\olf 4;23) 1 |O) | x] 5 5) 16} 5 | K| fa Number of assignments = order of matrix = 5. <. This is an optimal solution. We have assignment as, 13B, 35E,5>D,23A,43C :. Maximum profit = 38+ 40 +374 41435 191 Rs.Unbalanced Assignment Problem 3.5 Whenever the number of jobs are not equal to the number of facilities i.e. the cost matrix is not a square ‘matrix then such assignment problems is called as unbalanced assignment problem. To make such unbalanced assignment problem balanced, the cost matrix has to be made a square matrix. This is done by adding dummy rows or columns. The costs in such dummy rows or column are taken zero. In this way we convert unbalanced assignment problem into balanced assignment problem.Jobs Machine Mi Me 1yi0 aw 2/6 | 3Ll5 2 qhis is an unbalanced assignment problem. We add dummy column to this matrix with zero costs to make is problem balanced. Hence we have Here Ms is dummy machine ») This is also un! ‘We make this unbal: rith zero costs. Hence we have job Jobs Jobs ‘Machine M: Mz Ms 10 un Oo 6 3 0 5 2 (0. ‘Machine M: M2 Ms als 3 il wee palanced assignment, since number of jobs are not equal to number of machines. anced assignment problem as balanced assignment problem, by adding one dummy job 53 M: Mz Ms hpi 2 3 h/8 9 10 BLO 0 0—__ os “&- 29-3 IN a particular plant there are 4 machines to be installed. There are 5 vacant places available. The cost of installation of machines at different vacant places are given in the following table. Places) A | B | C | D| E Machines My 9 |u|is {10 | a M, 12| 9 | 10] 15| 9 M; w [Blu u M, w|eafliz|7|s Find the optimum assignment. Soln. : In the given A. P. Number of Machines # Number of vacant places AP. is unbalance, to make it balanced by adding dummy row with zero cost. A|Bplic|plxE M, | 9 [11 | 15 | 10 [a1 M: | 12] 9 | 10|15| 9 Ms | 10 | 13 | 14] 11 | 7 M,| 14] 8 [12] 7/8 Ms|o0lo}olojo Step 1 Row wise ATB[C [DE M/f)|2}6]1)2 M3 |} 1)6| «8 M)3/6/7/) 4 | Mi} 7] 1] 5 |] a Ms} @ | a |[] | a| aomultiple Optimum Solution (Alternative Solution) problems for @ final reduced matrix, There are two or more ways to make assignments In such ss Pe alternate opsimal solutions wih same optimum cost. j examples CxS te folowing assignment problem for minimum cost. pe loamy AT2 3 45 B/4 5 67 Cc 789 8 ols 5 84 gin: Given cost matrix is Iu miv Af2 3 4 & eee c}/7 8 9 8 DL3 5 8 4 subtracting row minimum from each row element, we get redueed matrix as follows : In mi Aro 1 2 Bi/O 12 3 c}/o 121 pLo 2 5 1 Subtracting column minimum from each column element, we get reduced matrix as In mv A oo 0 2 Bio-.a 6 2 c ooo0 0 D 013 0 Now we make assignments as follows : ss that there is no row or column with single zero. Therefore we make assignment arbitrarily. Let us assignment to the cell. a wa ean ‘A and column I. Next make an assignment to the cell (B,II) and cross other yy, and column II. Also make an assignment to cell (C,III) and lastly we make assignment to the cellHence we have assignments as follows : 2) * * yvouwpD x Re * a * ‘The optimal solution is A3I,B51L,C>1,D31V Minimum cost = 2 + 5 + 9+ 4= 20. Likewise if we firstly make assignment in the cell (A,II) we get alternate optimal solution as follows = Io mW Vv «x GK 2 A pl fo) wm K 2 Cc} « xo K D x 1 3 fl A SILBoLC3Il,DoV Minimum cost = 3 + 4+ 9 +4 = 20 Rs.ion N= gana Problems with Restrictions : _ assignment P on such that a * gipficulty ca” be st spe possibility of ass t ent. om ned Examples particular job can not be assigned to a particular facility. ignment of such restricted element will be automatically excluded from the roblem there may be restrictions on the assignments of jobs to the facilities. ie. there colved by assigning a very high cost (2) so the corresponding element in the matrix, e optim: 7a save the follo ee machine A. Jobs Aewne In. Since there is restriction that job 1 is spent of the given matrix. Th Jobs unene Now we solve this by Hungarian method. Subtracting row minimum from each row elem whore > Rw vst Machines Bc OOD 4 7 3 2 6 3 65 7 a7 Go 4 5 7 Subtract btracting column minimum from each column element ewe wiMeam erefore, the cost matrix is Machines A BCDE o 47 3 4 4 2 6 3 4 7657 5 3.3 767 445 7 3 jent we get, A 4) oi 4 of AO dt a 7120 2 0 0043 4 112 4 0 t we get, BCODE eo 1 4 021 O42 a2 0 2 004 3 4 112 4 0 wing assignment problem for minimization where no assignment Cal 5 not assigned to machine A, we put a high cost n be of job 1 (ce) to the restricte—$ $$ ee We make assignment as follows : Lo} - AB CD eo ee 2 ole Ol a 3) 21 0) 2 4) fo] w 4 8 5 ee ee Number assignment equal to order of the matrix. ~. This is an optimal solution. Which is given by 13D,2>B,35C,454,55E and optimal cost = 3+2+5+3+3 = 16. ol = KY