Transportation Problem “Commodities are transported while people travel” 10:1. INTRODUCTION The Transportation Problem is one ofthe subclasses of L.P.Ps. in which the objective is to transport various quantities of a single homogeneous commodity, that are initially stored at various sources (origins), to different destinations in such a way that the total transportation cost is a minimum. To achieve this objective we must know the amount and location of available supplies and the quantities demanded. In addition, we must know the costs that result from transporting one unit of commodity from various origins to various destinations. 10:2, LP FORMULATION OF THE TRANSPORTATION PROBLEM Consider the following transportation problem as an L.P.P. : A company manufacturing air-coolers has two plants located at Hyderabad and Mumbai with a capacity of 300 units and 100 units per week respectively. The company supplies the air-coolers to its four showrooms situated at Bangalore, Chennai, Delhi and Ernakulam which have a maximum demand of 85, 150, 150 and 55 units respectively. Due to the differences in raw material cost and transportation cost, the profit per unit in rupees differs which is shown in the table below : Gas Bangalore Chennai Delhi Emakulam Hyderabad 110 90 75 35 Mumbai 65 50 80 45 Plan the production programme so as to maximize the profit. The company may have its Production capacity at both plants partly or wholly unused. Formulate the above problem as an L.P.P. Mathematical Formulation Lets, represent the quantity of air-coolers to be transported from Hyderabad and Mumbai to Bangalore, Chennai, Delhi and Ernakulam respectively. The objective function is : Maximize z= 11Qr, + 90t2 + 75x13 + 55x14 + 65x) + SOry + 80x; + 45x54 The constraints are : (@) My + x2 +13 + ta = 300, An + X22 + 23 + X24 = 1005, 5, x12 + ap = 150, m3 + 493 = 150, ayq + pq = 555 © aye 2 and f= 1,234. wnt the above LiP.P. there are mx = 2x4 =-8 decision variables, x, and m+n = 6 constraints, ere m denotes the number of rows and n the number of columns. © yeas Scanned with CamScanner248 OPERATIONS ese, In general, let there be m sources of supply, SE = 1 2, 2, m) having a,(i = 1, 2, Of supply to be transported among n destinations D,(j = 1, 2, + m) with by (J = 1, 2, of demand. Also let cy = cost of transporting one unit of commodity from source i to dex and x, = quantity transported from source i to destination je Then, the problem is to determine the transportation schedule so as to minimi transportation cost satisfying supply and demand constraints, Mathematically, the problem may be stated as a linear programming. problem as follows - Minimize z= 2 & x,c, subject to the constraints : 1a +) ig *) ahen | the tng a A (Supply constring E xy (Demand constrains, and 2y 20 for all i and j Remark, Since all xy 2 0, it follows that a; 2 0 and by 2 0. Moreover, any ag = 0 implies tax x = O and thus can be eliminated from the problem. Therefore by requiring a; > 0 and by > 0, we ae not restricting anything, 10:3, EXISTENCE OF SOLUTION IN T.P. The general transportation problem always have a solution. In this section, we shall discuss #e theorems on the existence of feasible solution, optimum solution and basic feasible solution Theorem 10-1 (Existence of Feasible Solution). A necessary and sufficient condition forte existence of a feasible solution 10 the general transportation problem is that Total Supply = Total Demand , i.e., 5 a =f, bj = (say). mt Proof. The condition is necessary. Let there exist a feasible solution to the T.P. Then, we kre £ a, =f, by = A (say) in Sufficiency. We assert that there exists a feasible solution given by x, = a,b,/2 for all i anti Clearly, x, 2 0 since a; > 0, bj > 0 for all i and j. * 2 a 2 2-8 Also, Eve Bb M =F Ey = ” ” bom jez! and 2 2 Mat Ea = by Miele les Thus, xj satisfies all the constraints of the T.P. and hence is a feasible solution. Corollary. (Existence of an Optimum Solution). There always exists an optimum solation 2 re Proof, Let Ea, = % by, so that a feasible solution Ay exists. It follows from the constraints problem that each x, is bounded, viz, OS xy S min. (a, bp Scanned with CamScanneros nsPORTATION PROBLEM 249 1" ‘Thus the feasible region of the problem is closed, bounded and non-empty and hence there exists optimum solution, sno? theorem 10-2 (Basic Feasible Solution). The number of basic (decision) variables of the general nation problem at any stage of feasible solution must be m+ n~ 1. an Proof. Consider the m +n constraints of the transportation problem : X= and Awe Gm bce ms Oe be Dem) ‘Taking summation on both sides, these yield = = x BOE ant Subtracting the two, we have ; acl om m 2 met o a ea "ty - £ a 1h Sa-d Swclho- Be " (- Be $5) " which happens to be the last constraint of the problem. This indicates that if the first m4n—1 constraints are satisfied, then Za, = Eb, ensures that the a) (m4njth constraint will be automatically satisfied, Thus, out of m+n equations, we have only (m+n 1) linearly independent equations. Therefore, a basic feasible solution will consist of at most (m+n-~ 1) positive variables, the rest being zero. Further a feasible solution involving exactly (m +n-~ 1) positive variables is known as non-degenerate hasic feasible solution, otherwise it is said to be degenerate basic feasible. Remarks 1. When the total demand is equal to total supply, the transportation problem is said to be balanced and otherwise unbalanced. 2 The allocated cells in the wansportation table will be called occupied cells and empty cells will be called non-oceupied cells. 10:4, DUALITY IN TRANSPORTATION PROBLEM Since a transportation problem is itself a linear programming problem, there always exist another fased upon the same data and having the same solution. Let the original linear programming problem b: : i ansportation problem be the primal problem and its associated one be its dual problem. For the general transportation problem, the primal-dual pair is : Primal Problem. Consider the general transportation problem as Minimize z= EExy¢y subject to the constraints = vs xj)20 — forallé=1,2,.... mand j= 1,2, Eaxy Dual Problem. Since the primal problem is in its standard form, dual problem will be as follows : sf = Euja, +E 1b; subject to the constrains + a, Ey = Fay Maximize ay + Sey for all i= 1,2... mand j = 1,2,. ty, and v, unrestricted in sign for alli j Where uC = 1, 2, ceag m) and v,(j = 1s 2s om nm) are dual variables. Scanned with CamScanner. OPERATIONS | PESEARG, | The dual solution of the transportation problem implicitly measures the comparative . ~ kk ‘antage of the sites of factories (source of capacities) and the delivered price of the Produc! 250 adv: 10:5. THE TRANSPORTATION TABLE Since the T.P. is just a special case of general L.P-P.. the application of Simplex method woul, doubt. give an optimum solution to the problem. However. fortunately. simplex like methog solving such problems has been developed. Whenever it is Possible to place the given LPP, a transportation framework, it is far simpler to solve it by Transportation Method than by the si method. To facilitate presentation and solution, the transportation problem is generally portayey et tabular form as shown below . ‘ Destination Demand Table 10.1 The mn large squares are called the cells. The per unit cost c, of transporting from the ith orga 01 t0 the jth destination D, is displayed in the lower right position of the (i, jth cell. Any fesse solution to the T.P. is displayed in the table by variable The various origin capacities and destination requirements are listed in the right most (outer) colin 4y at the upper left position of the (i, jth cell and the bottom (outer) row respectively. These are called rim requirements. Remark. In order to avoid difficulty to discriminate between degenerate and non-degenerate basic solutions, we shall adopt the convention that while displaying a basic feasible solution in & transportation table, the value of any particular xy is actually entered if, and only ify ty is 4 BOSE variable. 10:6. LOOPS IN TRANSPORTATION TABLES a Definition (Loop). In a transportation table, an ordered set of four or more cells is said to form LOOP if (i) any two adjacent cells in the ordered set lie either in the same row or it oe column, and (ii) any three or more adjacent cells in the ordered set do not lie in the same ree same column, The first cell of the set is considered 10 follow the last in the set, ie, each cell the last) must appear only once in the ordered set. _ Consider the following two cases represented in Table 10.2(a) and 10.2(). In Table 10200 join the positive allocations by horizontal and vertical lines, then a closed loop is obtained. ordered set of cells forming a loop is L={G. 2), 2 2. 2 3). U1, 3) (1. 5.8, 5)} Scanned with CamScannerN PROBLEM 251 ATATION ansPo} ‘re loop in Table 10.2(b) is not allowed, because it does not satisfy the conditions of a loop- i + - LI. 4 _] on Table 10.2(a) Table 10.2(6) ‘Remarks. 1. Every loop has an even number of cells. 2, The allocations are said to be in independent positions if day independent individual allocation without changing the positions of these allocations, the rim conditions. That is, a closed loop cannot be formed through these allocations. + gach row and column in the transportation table should have only one plus and minus sign. All cells dia have a plus or minus sign, except the starting unoccupied (non-basic) cell are occupied (basic) calls. 4, Closed loops may or may not be square in shape. Definition 2 (Set containing a loop). A set X of cells of a transportation table is said 10 contain a Loop if the cells of X or of a subset of X can be ordered (sequenced) so as to form a loop. Theorem 10-3 (Linear Dependence and Loops). Let X be a set of column vectors of the cuffcient matrix of a TP. Then, a necessary and sufficient condition for vectors in X to be linearly dependent is that the set of their corresponding cells in the transportation table contains a loop. Proof. Exercise for the reader. Corollary. A feasible solution to a T-P. is basic if, and only transportation table do not contain a loop. Proof. Exercise for the reader. Remark. The above corollary provides us a method to check graphically whether the current feasible solution to the T.P. is basic or not. is not possible to increase or decrease or violating if, the’ corresponding cells in the 10:7, TRIANGULAR BASIS IN A T.P. The reader may recall that a system of ” linear equ: matrix A is triangular, ie., it is of the form ations Ax = b is called a triangular system if the ayy ayn 3 0 a2 43 a2_ A=| 0 0 ay ay, |, jy #0 for each i. 600 «0 Ayn fe is at least one equation having only one variable. This ‘aining equations, resulting in a reduced system of (n-1) a triangular system can be easily ‘solved by vag lhe system has a property that ther C ‘able can be eliminated from the rem ‘uations, which again happens to be triangular. Thus, substitution’. a ybefinition (Triangular basis). A basis for the system Ax = b is said to be a Triangular Basis if angular system is obtained when all the non-basic variables are set equal to zero in the system. Scanned with CamScannerOPERATIONS PEE, Theorem 10-4 All the dasis for a transportation problem are triangular, Proof, Let the set of constraints of a TP. be: Sayer PLA ome and abe FR wm pee Suppose that these constraints have been displayed in the transportation table having m Tows columns, We shall prove that : oo) () each row and colunmn of the transportation table contains at least one basic variab) (2) i is mot possible that all the rows and columns have Nvo oF more basic variables, ‘The result () follows from the fact that if any row or column does not contain then the corresponding rim requirement cannot be satisfied (as a's and bys Prove result (#) by contradiction, If possible, let us suppose that all the rows least two basic variables, le, and Na basic vay are all non-zero) ie and columns hae y ‘Then since there are mtn equations, the total number of m+n Therefore, i N2 Ww Three cases now basic variables will be at kay N denotes the total number of basic variables, we must have N's 2m and ise : Wi m>n then N22 3 Nemen Win>m then N22n => Nonem If m=a, then N22m => Nentm Thus, we observe that in every case V2 m+n. But, as we already know, N = men= 1, whic 'S a contradiction, Thus, our assumption about the existence of at least two basic variables ih Constraint equation is wrong. Combining (i) and (ji) we, therefore, conclude that there is at lead oe constraint equation that contains exactly one basic variable, We can eliminate this basic variable from all the other constraints, This results in a redved Seacboration table, with one row or column deleted and rim requirements readjusted, Repeating te above argument for this reduced transportation table we conclude that there is at least one tow (ot column) that contains exactly one basic variable in the reduced system. Eliminating this basic varibe from the system and repeating the above argument for the new reduced system we see that the rev | | | again holds for this reduced system. Repeating the above argument again successively, we see that the system is completely solved 10:8. SOLUTION OF A TRANSPORTATION PROBLEM The solution to a transportation problem involves the followi Step 1. Formulate the given problem as a linear progra ‘Step 2. Set up the given L.P.P. in the tabular form known as a transportation table. yo Step 3. Examine whether total supply equals total demand, If not, introduce a dummy roweo} i having all its cost elements as zero and Supply/Demand as the (4ve) difference of Supply ® Demand. Step 4, Find an ii conditions. improve Step 5. Examine the solution obtained in step 4 for optimality, i.e., examine whether an i transportation schedule with lower cost is possible. cout ‘Step 6. If the solution is not optimum, modify the shipping schedule by including that u™ cell whose inclusion may result in an improved solution. Step 7. Repeat step 4 until no further improvement is possible. We shall now discuss various methods available then attaining an optimum solution. ing major steps : mming problem. mat tial basic feasible solution that must satisfy all the supply and tion ic feasible sol for finding an initial basic feasible s° Scanned with CamScanner> arTiOn PROBLEM 253 1 * pINDING AN INITIAL BASIC FEASIBLE SOLUTION 109" eral methods available to obtain an we ee he following three methods = pl West Corner Method, Nomicost Method, and 2, 8S approximation Method (or Penalty Method). al basic feasible solution. However. we shall 5. vose! ip-West Corner Method (NWC Rute) North “simple and efficient method to obtain an initial basic feasible solution. Various steps of the a are: : eats 1, Select the north-west (upper left hand) comer cell of the transportation table and allocate fas possible so that either the capacity of the first row is exhausted or the destination 8 oven of the first column is satisfied, ie., x4, = min. (a, by). i 2 If by > a We move down vertically to the second row and make the second allocation of sup cityde 3, = Min. (4a, by —¥y)) in the cell (2, 1). jf, < a. We move right horizontally to the second column and make the second allocation of sagntude Xp = min. (4, —Xjy5 3) in the cell (1, 2). “If'b, = dy there is a tie for the second allocation. One can make the second allocation of site, M X12 = min. (a, — a1, ;) = 0 in the cell (1, 2), fa 4X, = min. (ay, by — by) = 0 in the cell 2, 1). Sep 3. Repeat steps I and 2 moving down towards the lower/right comer of the transportation se until all the rim requirements are satisfied. SAMPLE PROBLEM 1001. Obrain an initial basic feasible solution 10 the following transportation problem using the torh-west comer rule : D E E G Available 4 m 13 7 14 250 a 16 18 14 10 300 c 21 24 13 10 400 Requirement 200 25 275, 250 [Pune M.B.A. 1999; Madras B.Com. 2005] Solution. Since Za, = 5b, = 950, there exists a feasible solution tothe transportation problem. i i bain intial feasible solution as follows : Seng (MSPoRation table of the given problem has 12 cells. Following north-west comer method, “allocation is made in the cell (I, 1), the magnitude being x,, = min. (250, 200) = 200. The allocation is made in the cell (1, 2) and the magnitude of the allocation is given by Te tn, xy2 = min. (250 - 200, 225) = 50. ee allocation is made in the cell (2, 2), the magnitude being x2 = min. (300, 225 —50) = 175, i Tea of fourth allocation in the cell (2, 3) is given by xy = min. (300-175, 275) = 125, The Ses 4 on is made in the cell (3, 3), the magnitude being x,, = min. (400, 275-125) = 150 and tee ie fast) Allocation is made in the cell (3, 4) with magnitude x5,'= min. (400-150, 250) = 250, ‘Se 193, "tial basic feasible solution to the given T.P. has been obtained and is displayed in Scanned with CamScanner254 OPERATIONS p; EStaRg, 2000-22250 —27S 250 Table 10.3 ‘The transportation cost according to the above route is given by 5 = 20K IL + SOXI3 + 175K 18 + 125K 14 + 150 13 + 250% 10 = 12,200, Remarks 1. The cells which get allocation will be called basic cells. 2 [The initial basic feasible solution obtained by means of north-west comer rule may be far from optimum, because the costs were completely ignored, 2. Least-Cost Method or Matrix Minima Method This method takes into account the minimum unit cost and can be summarized as follows : Step J. Determine the smallest cost in the cost matrix of the transportation table. Let it be Allocate x, = min. (a, b) in the cell (i, j). Step 2. Ixy = a, cross off the ith row of the transportation table and decrease by by a Goto sp Ixy = b, cross off the jth column of the transportation table and decrease a by by, Go to Sup 4 If xy = a, = b, cross off either the ith row or jth column but not both, Step 3. Repeat steps J and 2 for the resulting reduced transportation table until all the ia {rauirements are satisfied. Whenever the minimum cost is not unique, make an arbitrary choice aot the minima, SAMPLE PROBLEM 1002. Obtain an initial basic feasible solution to the following T.P. using the matrix mitt method : D>, Ds Dy Dy Capac O 1 2 3 4 e O; 4 3 2 0 f Os ° 2 2 1 10 Demand 4 6 8 6 where O; and D, denote ith origin and jth destination respectively. Solution, The transportation table of the given TP. has 12 cells. Following the least-cost fies the first allocation is made in the cells (3, 1), the magnitude being x,, = 4. This satis! Tequirement at destination D, and thus we cross off the first column from the table. The allocation is made in the cell (2, 4) of magnitude x3, = of the table. This yields Table 10.4. There is, again, a tie for the third allocation. arbitrarily the cell (1, 2) and allocate x12 = min. (6, 6) = 6, there. Cross off either the or the first row. We choose to cross off the first row of the table. The next allocation = 0 is made in cell (3, 2). Cross off the second column getting Table 10.5. se it Xs; Scanned with CamScanner255 ( 6 8 Table 10.4 Table 10.5 choose arbitrarily again, 10 make the next allocation in cell (2, 3) of magnitude (2, §) = 2. Cross off the second row. This gives Table 10.6. The last allocation of = min, (6, 6) = 6 is made in the cell (3, 3). we min. rite A oy Table 10.6 Table 10.7 Xow, all the rim requirements have been satisfied and hence an initial feasible solution has been faemined. This solution is displayed in transportation Table 10.7. | Since the cells do not form a loop, the solution is basic one. Moreover the solution is degenerate 4lo The transportation cost according to the above route is given by 6x24 2xK24+6x0+4x0+2xe + 6x2 = 28+ 26 = Base +0. 4 Vogel's Approximation Method (VAM) Te Vogel's Approximation Method takes into account not only the least cost cy but also the costs that Mtaceed cy, The steps of the method are given below : Sep 1. For each row of the transportation table identify the smallest and the next-to-smallest «ah Desemnine the difference between them for each row. Display them alongside the transportation *y enclosing them in parenthesis against the respective rows. Similarly, compute the differences ‘exh column, tat? 2 Wdentify the row or column with the largest difference among all the rows and columns, tg US Use any arbitrary tie-breaking choice. Let the greatest difference correspond to ith row tage be the smallest cost in the ith row. Allocate the maximum feasible amount Teng,” > 5) im the (i, jth cell and cross off either the ith row or the jth column in the usual Sq wrk 2. Recompute the column and row differences for the reduced transportation table and g0 to “Peat the procedure until all the rim requirements are satisfied. nade A "OW or column “alfference” indicates the minimum unit penalty incurred by failing to 2 iy cation to the smallest cost cell in that row or column. Scop. sen later that VAM determines an initial basic feasible solution which is very close to "ths aoe’ ‘lution, that is, the number of iterations required to reach the optimum solution is smaller Scanned with CamScanner