The Assignment Problem, 13 Routing Problems | MSs "OR. produces an integrated solution for the good of entire organisation and not just to effect local improvement" 7:1. INTRODUCTION | ‘An assignment problem is a particular case of transportation problem in | which a number of operations are to be assigned to an equal number of operators, where each operator performs only one operation. The objective is to taaximize overall profit or minimize overall cost for a given” assignment schedule. ‘An assignment problem is a completely-degenerate form of a transportation problem. The units available at each origin and units demanded at each destination are all equal to one, That means exactly one occupied cell in each row and each column of the transportation table, i.e., only ‘occupied cells in } place of the required n +n —1(=2n—1). | 7:2 MATHEMATICAL FORMULATION OF AN A.P. Consider an assignment problem of assigning n jobs (operations) to» persons (operators). Let cy be the cost incurred in assigning ith operation to jth ‘operator and let : fi if the ith operation is assigned to jth operator = 10 otherwise Xi ‘Then the assignment problem is simply the following LPP : Minimize 2= LD xyey asa jet subject to the constraints : Lys eet and with xy=0, or 1. An AP, could thus be solved by simplex method. It also happens bean nx n transportation problem with each a; = bj = 1. ‘However, as an assignmentsimp!2x method or transportation meth OPERATIONS RESEARCH 184 ting to attempt to solve it by problem is highly degenerate, it will be frastrain B 10 convenient iterative solving an AP. m assignment sch from all the eleme edule remains unaltered if procedure is available for nis of the row or column of Theorem 7-1. The optimu we add or subtract a constant (0! i, it matrix. ‘ ea we ciate assignment have cost matrix ¢ = (Cy). # aaa an * for all i and j, Let cif = cy mY to/from all the elements of the ‘That is, we add or subtract a constant ij, Yj ith row and jth column of ¢ respective objective function is sett Ty. Then for the new cost matrix e*, the z j (cy twit Y; J Rwcosnty i Suet d duet Bh eer =2tUtV =bmved vi Bape Exel where 0 = uy Va Dy (. tu = 2 49 Since’ U and V are constants, an assignment (x;) which minimizes z will also ypinimize 2°. heorem 7-2. If for an assignment problem all cy 2 0, then an assignment schedule (xj) which satisfies LEX, cy = 0, must be optimal. Proof. Exercise for the reader. The above two results form the basis of Assignment, Algorithm. By selecting suitable constants to be added to or subtracted from the elements of the cost matrix we can ensure that each c,* > 0 and can produce at I i. 7 iy =! least one ¢y in each row and each column and try to make assignments from among. the: positions. The assignment schedule will be optimal if there is exactly one assignment (ce., exaclly one assigned 0) in each row and each column. 7:3. ASSIGNMENT ALGORITHM \&4¢ Various steps of the computatior mn . assignment may be summarized as, Token gerocedure for obtaining an optimal Step 1. Check whether the cost matrix is sq Step J ist matrix is square, i adding suitable number of dummy rows (or columns) wit ost Semen " Step 2. Locate the smallest cost i A “lest cost elements in each row of i trix. Sa nc geroin cach rower clement i that Tow, Asa result tere > Tow of the reduced cost matrix.THE ASSIGNMENT PROBLEM, ROUTING PROBLEMS Step 3. In the reduced cost matrix is “, ioe the smallest clement in it. Subtract the ano er each column and locate the column. As a result, there would beat least one-ner cn CvetY other entry in columns of the second reduced cost mattix “TiN cach of the rows and Step 4. In the above reduced as follows : () Examine the rows successivel found. Enrectangle this zero as Q) Cost matrix search for an optimum assignment / OF column of the reduced matrix, then arbitr: ily /, choose a row or column, having. the minimum number of 0's. Arbitrarily select and enrectangle any one 0 in the row or column thus chosen and cross efi steet O's in its tow and column: ‘Repeat steps (i) and. (ii il "« either assigned or crossed. PS (i) and. (i ) until all the 0's have been _ {if If each row and each column of the reduced matrix hag one and only one assigned 0, imum assignment is made in the cells of enrectangled 0's. Otherwise go to the next step, Step 5. Draw the minimum number of horizontal and/or vertical lines through all the 0's as follows : = (a) Mark (1) the fows}in which assignment hg not been made. (b) Mark (1) coluinns which tiave zeros in the marked rows. © Mark (1) x6ws (not already marked) which have assignments in marked éolumns, Sarasa Grae “@ Repeat (6) and (c) until the chain of marking is completed. (©) Draw straight lines through all @imarked rows} columns, Step 6. If the minimum number of Tines passing through all the zeros is ‘equal to the number of rows or columns, the optimum solution is attained by an arbitrary allocation in the positions of the zeros not erosse Otherwise go to next step. Step 7. Revise the costs gna follows 7 (® Find the smallest eletnenf not tovered by afiy of the lines of ste Gi) Sibtract this From all the w point of intersection of the two lin ~Gii) Other elements crossed by the lines én > aa Step 8. Go to Step 4 and repeat the procedure till an optimum solution is elements and-add)the same at the Ce Note : The above iterative method to determine an assignment schedule is known as Hungarian Assignment Method. | 3 \ ctl Remarks : 1. There are situations when a particular assignment may Net | be permissible. In such situations we asssign a very high cost (say ) | an assignment and proceed as usual. : *. 2, If the assignment problem involves maximization, cone ‘matrix to an opportunity loss matrix by subtracting each e/SMC % pea ¢ effective from the vit186 OPERATIONS RESEARCH i i ix i bh highest element of the matrix. Minimization of the resulting matrix 1s the same! as the maximization of the original matrix. | SAMPLE PROBLEMS : ur subordinates, and four tasks to be 701.:A departmental head has four ss inate vases differ in their performed. The subordinates differ in efficiency, varincie difficulty, His estimate, of the rime eack man would take (0 perform each task, is given in the matrix below : Men qT _ vasks ; ; a ¥ z 78 26 17 W B 13 28 14 26 c 38 19 18 15 D 19 26 24 10 to a man, so as to plinimize the total 1993; Bharathidasan (auied) 1992; (Math.) 1990} How should the tasks be allocated, one man-hours ? UAS 1990; Kerala M.Com. 1! Calicut B.Tech. 1990; Anantapur M. Solution.. Step I. Subtracting the smallest element of each row from every element of the corresponding row, we get the reduced mat 7 15\6 6 0 15 1 13 4 6 23044 3 0 9 16 14 0 Step 2. Subtracting the smallest element of each column of the reduced matrix from every element of the corresponding column, we get, the following ‘reduced matrix: vo \ y > 7 1 5 0 o 11 0 13 23 0 2 «0 9 12 13.0 Bee. 7 ca it aaa enrectangle C] (i.e.,. make assignment) 2 — RG Sl ote ree fn te column oe marked: Them, 7 us @ MO ue 23 2 9 12 130 & In the above matrix, we arbitraril row 2 had two zeros. NY enrectangled a zero in column J, because #a 8: {ASSIGNMENT PROBLEM, ROUTING PROBLEMS io qr may be noted that colurnn 3 and row 4 do not have any assignment. So, we we on 19 the MeXt SUP. _ a 14 ( Since fow 4Y06s nothave any assignment, we tick this row ( ) Step (i) Now there is 8 z6r0 Jin the fourth column of the ticked tow. So, we tick fousth eoturan ( °.)- — (if Further there is an assignmént in the first row of the ticked eolimn. So we tick fi tow (V). — — cee Draw straight lines through all unmarked rows and marked columns Thus, we have v v Step 5. In step 4, we observe that the minimum number of lines so drawn is vi 3, which is less than the order-pf the cost matrix, indicating that the current, assignment is not optimum. To increase the minimum number of lines, we generate new 25°05 in the erate new 26905 modified ifatrix. chef ad ° ‘Step 6. The smallest glement not covered by she lines is xs g this ithe same to all the elements element fom al]-the Zed elements and lying at the dor of the Jines , we obtaifr the following new reduced cost matrix a 56 0) 0 o 1 0 18 230 2 5 4 7 8 0 Step 7. Repeating step 4 on the teduced matrix, we get 2 6 ca @ uw & 18 23 OM 2 5 47 8 Now, since each row and each column has one_and i optimal solution is reached. The optimum tasignment is saly ope assignments © A+G,B+E, CF md DH. ‘The minimum total time for this assi is i ae ae for this assignment scheduled is 17 + 13 + 19 + 10OPERATIONS RESEARG 188 , ind 5 districts. Consideri manager has 5 salesmen ar sidering 4 702 A erie categmen and the nature of districts, the marketing manage, capabilities o} th (in hundred rupees) for each salesman in each estimates that sales per mon Machine Job A B C D = ———7 38 40—«H 40 24 28 21 36 4102733 3037 22 38 Al 26 36 29 «330 40385893. BAW KA Sry TT TFT Find the assigament of salesmen to districts that will result in plaximuh sales. [Nagarjuna B.Tech, 1985; Bharthiar B.E. (Comp.( Se) 1945; * Bharathidasan B.Sc. (Math.) 1993, M.Sc. (Appl. Sc.) 1987} Solution. Convert the profit matrix into opportunity loss matrix by subtracting all thé entries from the highest element 41 of the given inatrix. Initial Iteration. Reduce the opportunity loss matrix so that there is at least ‘one zero in each row and each-column. Make the proper assignments in rows and columns. Also draw the minimum number of lines to cover all the zeros of the reduced matrix. See Table 7- oe nas den] Table 7-1 Final Iteration, hast Table 7-2 - Modify Table 7-1 by subtracting the element ‘4’ from all the elements not covered i : lines. Thus, we get “te, _ Hines and adding the same at the intersection of two ‘The optimum assignment is TI B2943 > B446,5 4 Ds or 1B BIAS Ds or (9B AI RAD SoC. 1B BINA Ds a. A ye teh 1. 9 we ve A department head hay es, the subordinates digger on eetaets 10, % performed and three The estimates of the time, eacksquB ASSIGNMENT PROBLEM, ROUTING PROBLEMS 189 subordinate would take to perform, is given below in the matrix. How should he allocate the tasks one t0 each man, so as to minimize the total man-hours ? Task eee eee eee 1 2 3 I 9 26 15 I 13 27 6 ml 35 20 15 Vv 18 30 20 UAS 1993; Dayalbagh B.Sc. (Engg.) 1989; M.Sc. (Math.) 1989) Solution. Since the problem is ‘unbalanced, we add a dummy column with all the entries as zero and use assignment methods for optimum solution. Now reduce the balanced cost matrix and make assignments in rows and columns having single zeros. Thus, we have : Table 7:3 The optimum assignment is 1 1,1 ~ 3, and IH — 2; while task IV should be assigned to a dummy man, ie, it remains to be done. The minimum time is 35 hours. 1) PROBLEMS 704. A fim plans to begin production of three new products. ‘They own three plants and wish to assign one new plant. The unit cost of producing iat plant jis c,, as given by the following matrix. Find the assignment that minimizes the total unit cost. Plant 10 8 12 Product | 18 6 14 ~ |6 4 2 (Gorakhpur M.Sc. (Stat.) 1986] 705. A company has,4 machines Gn which to do 3 jobs. Each job can be assigned to one and only one machine. ‘The cost of each job on each machine is given in the following table: een ere COOTER --eeEEEsutn eC Job Machine a B c D 1 18 24 28 32 2 \ 8 13 7 WD 3 10 oe 9 2 Determine the optimum assignment. [Gauhati MCA 1992; Osmania MSe. (Are Math), 19851