Part 1 Topics 7A Characters and Assumptions ofthe “ransprtalon Problem 72° Structure ofthe “Transportation Table 7.3. Transportation Algorithm 7.4 Obtaining the Initial Solution 7.5 Optimal Feasible Solution Learning Outcomes + Define a transportation problem and list the assumptions. + Formulate the problem ina trangportation matrix. + Demonstrate how to find the inte feasible solution. + Apply the stepping stone ‘method and the MODI method. Distribution Models Introduction e A special type of an LP problem in the area of physical distribution of goods and services coming from several supply locations that are to be delivered to demand centers is called distribution model. There are two main types of distribution problems: the transportation problem and the assignment problem. The structure of the transportation problem involves shipments from a number of sources to a number of destinations. Its objective is to determine the most efficient routes from the supply centers to the demand centers that would meet the required quantity of goods and services of all destination centers relative to the availability of limited supply. Finding the cheapest shipping schedule that will satisfy the demand without the violation of the supply constraints is the main goal of the transportation problem. The assignment problem, on the other hand, deals with situations wherein a given number of candidates are assigned to a number of positions. For example, workers are assigned to certain machines or tasks. As there are different expenses or profits for each match, the goal is to determine the optimal best one-to-one match. This chapter will discuss the transportation problem. UMIEIPSL COLLEGE creme SAR MAIER,68 In 1941, Frank Lauren Hitchcock, an American mathematician and physi Presented the simplest form of transportation model in his study “The Distribution of Product from Several Sources to Numerous Localities.” The transportation model further developed by Tjalling Charles Koopmans, a Dutch-American mathemati and economist. He wrote “Optimum Utilization of the Transportation System” in 194 ‘Then in 1953, Abraham Chames and William Wager Cooper, two American operat researchers, developed the stepping stone method. The modified distribution meth was developed in 1955. Lesson 7.1: Characteristics and Assumptions of the Transportation Problem The main objective of the transportation model is to determine the chea Toutes from the suppliers to the destinations. Figure 7.1 shows the relations between the suppliers and destinations. A transportation problem exhibits following characteristics and assumptions: . 1. Supply or Sources. Certain sources have limited available quantity ‘one commodity. Demand. The demand comes from several destinations, such warehouses, distribution centers, shops, eto. Quantities. The quantities available at each source and the demands each destination are constant. Shipping Cost. The shipping cost is based ‘on per unit cost of Commodity from the source to each destination. It is usually based the distance between the two points, Demand and supply quantities are presented in whole numbers. SUPPLY/SoURCES DESTINATION CENTERS Figure 7.1. Relationship between the Supply and Destinations MANAGEMENT SCIENCESample Problem: Mr. Danny Atienza is the owner of Cement Exchange, a cement composites manufacturer with three plant locations: Cement-bonded composites are important construction materials that are made of hydrated cement paste that binds wood or fibers to make precast for building components. Cement Exchange is currently a top manufacturer. of cement composites and has been an exclusive supplier to four ‘major distributors in Metro Manila. The total monthly purchase of each of the four distributors are 1,500, 2,000, 2,400, and 3,500 boxes. The four distributors may receive deliveries from any of the three manufacturing plants. However, there are limitations to the production capacity of each plant: Plant 1 - 3,000 Plant 2 2,700 Plant 3 ~ 3,700 Mr. Atienza would like to reduce the logistics expenses by proper scheduling of shipments to the four distributors. The company’s chief accountant submitted an ‘estimate of the per box/unit shipping cost from each plant to each distributor. The Shipping costs are in Philippine currency: Plants Distributors i 2 3 4 1 15 18 22. 26, 2 21 21 16 23 3 14 19 20 24 ~ Based on the avallable data, the owner decided to use the transportation ‘method to compute for the cheapest delivery schedule, ‘Cement Exchange's distribution problem when converted to LP will show: Decision Variables: %,.= number of boxes shipped from plant 4 to distributor 1 ‘%.= number of boxes shipped from plant 1 to distributor 2 %,= number of boxes shipped from plant 1 to distributor 3 %..= number of boxes shipped from plant 1 to distributor 4 % = number oftbaxes shipped from plant 2 to distributor 1 a= number of boxes shipped from plant 2 to distributor 2 CHAPTER 7 DISTRIBUTION MODELS PART1 6970 lumber of boxes shipped from plant 2 to distributor 3 X= number of boxes shipped from plant 2 to distributor 4 X,, = number of boxes shipped from plant 3 to distributor 4 = number of boxes shipped from plant 3 to distributor 2 X-q= Number of boxes shipped from plant 3'to distributor 3 X= number of boxes shipped from plant 3 to distributor 4 The goal of Mr. Atienza is to minimize the total shipping cost. The transportation Cost Is computed by multiplying’ the amount shipped over each route by the per- Unit shipping cost for that route, and these costs are summed up to get the total transportation cost. The objective function for the Cement Exchange. problem is expressed as: Objective Function: yi Minimum © = 18x, + 18x" + 22x,,+ 26x,,+ 21x;, + 25%, 16X,,+ 23x,,+ 14x, + 19K,, + 20%, 24x, The problem of the company involves two types of constraints: demand Constraints and supply constraints. For the demand constraints: A decision has to be made if shipping from plant 1, plant 2, plant 3, or a combination thereof should be made to distributor 1, which only requires 1,500 units. Thus, the corresponding constraint should be: X44 +X) +%, = 1,500 (requirements of distributor 1) Inthe same way, the rest of the constraints are presented as: Xa # Xz + Xp = 2,000 (requirements of distributor 2) Xp + Xp + %5 = 2,400 (requirements of distributor 3) Xiq + Xy + %Xq = 3,500 (requirements of distributor 4) Take note that the demand of the distributors is equal to the total capacity of Cement Exchange plants. Therefore, the supply constraints must indicate that the goods shipped will not exceed the plant capacity. To mathematically formulate the constraints, we have: Xu + Xia Xp +X,, = 8,000 (capacity of plant 1) Xoy + Keg Xag + Xa, = 2,700 (capacity of plant 2) Xa + Xaa + Xoq + 1g, = 9,700 (capacity of plant 3) MANAGEMENT SCIENCELesson 7.2: ‘Structure of the Transportation Problem ‘Transportation problems are presented in tabular form to make it easier to apply and understand the transportation algorithm (Table 7.1). The following are the rules on how te i fo construct transportation table: Each row in the transportation table contains the source of supply, while each column is for the demand or destination point. The total available supply is written on the right side of the table's main body. For Cement Exchange, the capacities of plants 1, 2, and 3 are 3,000, 2,700, and 3,700, respectively. The total requirements of each destination are given in the lowest row of the main body of the table. In the case of Cement Exchange, distributors 1, 2, 3, and 4 have demand requirements of 1,500, 2,000, 2,400, and 3,500, respectively, i Cells in the main body of the table should correspond to the decision variables. Cement exchange has 12 decision variables, . At the corner of each table cell, the “boxed-in’ value contains the Objective function coefficient of the decision variable. In the case of Cement Exchange, cell 1—1 indicates that the costs to ship from plant 4 to distributor 1 is 15 per unit, Table 7.1 Transportation Problem for Cement Exchange Aaitere Tralebliy Peg 3000 n os 2 2700 oe, * 3700 Xe X X x % 750 72000 3500 300 CHAPTER 7 DISTRIBUTION MODELS PART 74Lesson 7.3: Transportation Algorithm Figure 7.3 shows the process'for the search solving transportation problems. and evaluation of the algorithm for ‘new solution, : (Ciclo ef Figure 7.3. Transportation Method Algorithm Step 1. Arrange the data in transportation table, Step 2. Obtain an initial basic fea: methods: Northwest Comer Method, (or Penalty) Method. Step 3. Test the initial solution for optimality. Step 4. If it is riot yet optimal, deter mine @ New improved solution, Evaluate the Updated solution. Repeat step 3 until la final solution is achieved. 72 MANAGEMENT SCIENCELesson 7.4: Obtaining the Initial Solution The first step in the transportation method is to find an initial feasible solution (Considered the first iteration) that does not violate the constraints. There are three methods to obtain the initial solution to a transportation problem: Northwest Corner Method (NWC), Greedy Method or Minimum Cost Method (MCM), and Vogel's Approximation Method (VAM). The northwest comer method (NWO) is the most simple and logical method of finding the initial solution. The procedure begins by allocating units to the upper left- hand corner and ends in the lower right comer of the transportation table. The minimum cost method (MCM) is a systematized procedure that is easy to use and yields an initial solution that is close to the optimal solution in small problems. mn this method, the allocation is made to the cell with the most possible lowest cost or the highest profit in a maximization case. 3 Vogel's approximation method (VAM) is an algorithm that obtains the initial feasible solution by determining the “penalty cost" of not using the lowest cost route. This method results in an optimal or near optimal as an initial solution. ‘Solving the transportation problem of Cement Exchange starting with the initial feasible solution using the three methods: A. Northwest Corner Method (NWC) Step 1. Startin the upper left-hand comer cell (northwest corner) of the main body of the table by putting the largest value that satisfies the supply constraint and demand constraint. Se pe ae] 7 6 @ 300 Pint Plant a a 700 Pant a « 7 3700 ee 300 CHAPTER 7 DISTRIBUTION MODELS PART 1Step 2. Once the supply constraint is satisfied, go to step 3. If not, move one cell té the right in the same row. Place in that cell the largest value that does not exceed both supply and demand constraints. In doing so, you have to consider al cell values that were previously entered either in the row or column, Repeat this procedure if the supply constraint is not yet exhausted, r a a Cerpuer | Dengue Darga | “EY Plant T Ez 3a 1500 L® Plant * x EE] 2 w ‘Step 3. Once the entire supply in a particular row has been‘Used up, move down to the next cell located in the same column. Place in that cell the largest value that does Not exceed the row or the column that corresponds to the demand. Again, all values previously entered must be considered. ‘Supply “Avalabiiy Dsipior | Datiputer Disiuter Pent Tio LL san = | Pant: [# [2s 2] 0 baat Es Pans Ts Te or Requirements | —3500— | Step 4. Return to steps 2 and 3 until all rows and columns are used up. ‘Suppy —_ Dai Destiputer Disteior ® Plant a Pants @ ‘Avalabity | Requirements 74 MANAGEMENT SCIENCEStep 5. Compute the value of the objective function by multiplying the value in each cell by the objective function coefficient *boxed-in values.” Then, sum all the cells. This is not a necessary procedure, but it is helpful to determine how much improvement in the cost (or profit) is gained in every iteration. ‘Guanty Shipped si Teal Cost 1500 250) T5007 2000 30 2300 20 3500 “otal Transportation Cast B. Minimum Cost Method (MCM) Find the cell with the lowest unit cost. Should there be a tie, choose randomly. Place in this cell a value that exceeds neither supply nor demand constraint. Repeat this procedure until all supplies and demands have been allocated. Step 1. The lowest shipment cost is 14 at cell 3-1. Allocate 1,500 in that cell and adjust the “availability” and “requirements” total accordingly. ‘Avalabiiy 3000 2700 a 2200 CHAPTER7 DISTRIBUTION MODELS PART 1Step 2. Cell 2-3 with the objective coefficient of 16 is now the lowest. The amount 2,400 was allocated to the cell, and adjustments were made on the totals. 1 ee oop ge ae] PatT = E a] 30 Pant a = 6 a) ae 2400 00 Pant * @ Fy BP) ia L 2200 Requkemenis | ~-#55= | —z000 | Bane —|—a500 30 Step 3. Among the remaining cells, the lowest this time is cell 1-2. « t t i : ‘Supply pues Wwailbiiy | Dat Ds i sipaior | Detter | Oevigior | Deter ; Prank % 8 z =] Sa : 4000 : Prank a B rg 2 ae yf 2400 300 § PiantS te L* © 2 a fe kK | Requirements | —500— 2000, Ee 33500 9.400 5 Step 4. The last remaining cells to be allocated in the following order: Cell2-4 300 Cell3-4 2,200 Cell1-4 1,000 76 MANAGEMENT SCIENCEThis step cancels the remaining values in the supply row and demand column. Sapp Tray pape [Sap sar ae od 1.500 ao Step 5. Compute the objective function. | —Guantly Shipped 11500 2000 00 300 ©. Vogel’s Approximation Method (VAM) Step 1. Subtract the lowest and second lowest values in each row and column. Select the row and column with the largest opportunity cost (largest difference). Lowest Coat 18 16 24 16 19 14 4 4 10, ct 20, 16 2 2B Rows 2 and 3 have the highest opportunity cost at 5. We can choose any of the two values. For this example, we choose row 2. CHAPTER 7 DISTRIBUTION MODELS PART1 77Step 2. Insert 2,400 units in cell 2-3 since it is the cell wi ‘Adjust the amount of supply (or demand) accordingly ey Distibutor Distiputor ; Plant T ‘Step 1. Row 3 has the highest pportunity cost, Step 2. Insert 1,500 in cell 3-1. Fi ‘Supply [Faw Paz i Pants Requirements 78 MANAGEMENT SCIENCE‘Step 3. Column 1 has been removed from further consideration. Repeat step 1. ‘Step 1. Row 1 has the highest opportunity cost. RowiColunin “Rows Row? Row. ‘Step 2. Insert 2,000 in cell 1-2 Step 3. Column 3 has been removed. Repeat step 1 Step 1. The opportunity cost is 1. (Rewicetamn 2d Lowest Gost ‘Column 4 2 Step 2. Insert 300 in cell 24. CHAPTER 7 DISTRIBUTION MODELS PART1 79,‘Step 3. Row 2 has been removed. Repeat step 1. Step 2. The lowest cost in column 4 is 24. insert 2,200 in cell 3-4 Step 3. Row 3 ha 1,000 in cell 1-4. [Supply Prati 80 MANAGEMENT SCIENCELesson 7.5: Optimal Feasible Solution The next step after obtaining the initial solution in the transportation method is to evaluate the net contribution of the routes that were not used. In other words, we would like to determine how much will be the increase or decrease in the objective function if any of the unused cells willbe included in the route. There are two alternative procedures that are used for evaluating unused routes: the stepping stone method and modified distribution. The stepping stone method (SSM) is generally the easiest to visualize and understand. Its purpose is to determine the effect on the total shipping cost should cone unit of goods are to be delivered through the unused routes. The modified distribution method (MODI) is a more efficient procedure in determining the net contribution of routes that were not used. In this method, the focus is on the extensive use of the cost factors associated with each cell . A. Stepping Stone Method (SSM) Step 1. Evaluate the net contribution of unused routes To illustrate this method, we shall use the initial solution under the NWC. Let Us consider the unused cell 1-3 (Transportation Tableau 1). An addition of one unit to that cell would mean that Cement Exchange is shipping 3,001 units from plant 1. This means that plant 1 will exceed its capacity. To offset the overcapacity, there has to be a decrease by one unit to either distributor 1 or distributor 2, Let us choose to reduce the shipment to distributor 2 at cell 1-2. However, this adjustment will make distributor 2 to receive less of the required 2,000 units. To solve this, we have to add another unit to cell 2-2. As a result, plant 2 is now one unit more than its capacity of 2,700. We need to reduce to the shipment to route cell 2-3 to correct the overcapacity of plant 2. The last adjustment has completed the chain of adjustments. Let us now ‘compute the net contribution of adding one unit to route or cell 1-3. CHAPTER 7. DISTRIBUTION MODELS PART 1 8Cell to be evaluated: Cell 1-3 Transportation Tableau 1 Supply ‘Avaiabiiy Disepaae 00 [5 Pant a 2700 PantT 5000 3700 An additional P13 will not contribute to the goal of lowering the total transportation Cost. Therefore, we must determine the net contributions of all unused cells. But before that, here are some rules in evaluating routes. * Start with the cell to be evaluated and trace a closed path through other ceils until you are back 10 the origina) an! being evaluated’ 73 make the closed Patt, itis allowed to skip overused or unused cells. Making a tum on a closed path is only allowed on used squares. There will be a single closed path to a Correctly formulated transportation problem. Assign plus sign or a minus sign to each cel thats part ofthe route, beginning with the plus sign on the cell being evaluated. Do the plus and minus signe alternately. Add a unit to a esl being added (with a plus sign) and subtract a unit to a cell with @ minus sign. Compute the net contribution of the chain of adjustments by adding per unit costs to all routes with a plus sign and subtracting per Unit Costs of all routes that contain a minus sign. The result may either be a Positive, negative, or zero contribution. Let us continue to determine the net Contributions for the other unused routes. 82 MANAGEMENT SCIENCECell to be evaluated: Cell 1-4 [Supply PaantT Piant2— CHAPTER 7 DISTRIBUTION MODELS PART 1Cell to be evaluated: Cell 2-1 Soap oa ges ope ii 3 a = Patio osemclulaa tle ss 7000 [Part ft lene ‘ 7 CEs Elie a] am Plant 3 4 1 2 4 ‘3,700 ey mE can Requremenis | 130 —|—2aa5 2A 3500 3700 Cell to be evaluated: Cell 3-1 ‘Supply Distributors ‘Availat Dietputor Seer, Behutor Detiutor nr [Pa a z =] aa Plant Z te a Me 2 +] 16 23 2,700 —»> er Plant 3 +] 4 19 YT 20 Py ‘3,700 : mT sso Reuremani | i500 | —a069 |} 2350 —| 58 —] eae 64 MANAGEMENT SCIENCEStep 2. Test for optimality For a minimization problem, all unused cells with negative contribution values will be favorable since they will reduce the transportation cost. Positive net contribution, on the other hand, will not be desirable as it will represent cost increases. The table below shows that there are negative contribution values in this problem. We should, therefore, continue the transportation method and proceed to step 3. The total ‘transportation cost for the initial solution is ®185,200. ‘Supply va [* 1500 me E @ Plant t Step 3. Select the new route The third step in the transportation method is done by simply selecting the route with the smallest or most negative net contribution for minimization cases. If there is atie, then select either route. For maximization problems, the route with the largest value must be selected. The Cement Exchange problem is a minimization problem. Thus, we select the route to cell 3-1 (with net contribution value at -12) for the next iteration, Step 4. Determine the quantity of goods to ship over the selected route |n the selected route, positive signs were assigned to cell 3-1, cell 2-3, and cell 1-2. These cells are called positive corners. The route with negative signs, on the other hand, are called negative corners. The highest value that can be assigned to cell 3-1 must be equal to the value of the negative comer of the closed path that thas the lowest value. CHAPTER7 DISTRIBUTION MODELS PART1 a5Maingetnal ants cece Qauuatea The smallest value among the negative corners is 200. ee, Distibitor Diatibaor re} oe a : Seer 2) sae | 30 —] ras Step 5. Develop a new solution (2nd iteration) For the new solution, the positive comers of the closed path will have additional 200 units, while negative.corners will be reduced by 200 units, The pew values to be entered in the new solution are as follows: Lowest Negative Comer Positive New Corners Values Cell 3-41 200 Cell 2-3 2,400 [Coiti-2 1,700 Negative New Corners Values Cell 1-1 4,900 Cali 2-2 300 Cell 3-3 0 Below is the transportation table for the new solution (2nd iteration) with the net contribution value of 182,800, which is lower by 2,400 over the cost of the initial solution, hey Pat |e MANAGEMENT sci HENCEStep 2. There are two negative contribution values (cell 2-1 and cell 2-4). This means we will continue to step 3, Step.3. The new route will be at Cell 2-4 as it has the lowest net contribution value. CHAPTER DISTRIBUTION MODELS PART1 47Step 4. The lowest value among negative corners is 300, which will be shipped to new route cell 2-4. ‘Supply Avallabity | The | Lowest Negative Comer 180,100, find out ift net contrib’ The t Positive Corners | Old Values Add New Values every unus Cell 1-2 1,700 300 2,000 ship 1,000 Cell 3-1 200 300 500 to distribut distributor Negative Corners | Old Values | Subtract | New Values to distribut Cell 1-1 1,300 300 1,000 Cell 2-2 300 300 0 Cell 3-4 300 3,200 cr raga | ‘Availabilty “> ve a, 26 3,000 Step 5. Develop a new solution (3rd iteration) 4 : i t z ic é “| : E 3 2 é @ BY] 2700 2 m%] 3700 3400) 88 MANAGEMENT SCIENCERAN MATER anHOInA! ner ee B. Modified Distribution (MOD!) Method. |n obtaining the optimal solution for SSM, the algorithm procedure includes determining the net contribution of shipping routes that were not used, The MODI method provides a more efficient way of computing the net contribution of unused Toutes. We shall use the Cement Exchange problem again and start from the initial solution that we have computed using NWC. Table 7.2 is the transportation table structure that we will use for MODI solution. Anew column on the left has been added tothe table, and we will let V,, V,, and V, be the new row values. For the new column values, we will be using W,, W,, and W,. Thus, we will have the following terms: V, = value assigned to row i W,= value assigned to column j Table 7.2 Transportation Table for MODI Method Pint Fy =| 7 eT | The cost of shipping one unit on a given route is the main focus of the MODI method. Used routes are utilized to solve for each V, and W, value. These values are Used to calculate the net contribution of unused cells. Let us not compute the final Solution to the transportation of Cement Exchange, Table 7.2 carries the solution that was obtained by NWC. Note that for every used cell, the relationship is expressed as: MANAGEMENT SCIENCE‘As we can see in Table 7.2, there are six used squares, Therefore, we can have the following relationships: 1.V,+W, =15 2.V, +W, = 18 3. V, +W, =25 4.V,+W,=16 5.V,+W,=20 6.V,+W,=24 Ait transportation problems have one redundant constraint, so we can select any of the values above. To make it easier, we can choose V, and make its value equal to zero. Thus, the values of W, and W, in equations 1 and 2 are 15 and 18, respectively. To solve for the value of V, in equation 3: e V, + W, = 25 . V, = 25-18 N,=7 Then, the same procedure will be applied to the rest of the equations to obtain all missing values. We now have the following values: v,=0 Using the expression below, let us now obtain the net contribution values of unused routes: \ NC=C,-V,-W, CHAPTER 7 DISTRIBUTION MODELS PART 1Table 7.3 shows the calculations to obtain the net contribution values: Table 7.3 MODI Method for Calculating Net Contributions CellRoute | C, | -V, Ceii-3 | 22 0 Call 1-4" | 26 0 13. Cell 2-4 24 i 15 cai24 | 23 7 13 Cell 3-41 14 it 15 “12 Cell 3-2 19 1 18 [10 Aiter calculating the net contribution values, we proceed to the steps in evaluating optimality and so on. These steps are repeated until an optimal solution is reached, ? Conclusion The advantage of using the transportation method is to help identify optimal transportation routes and the quantity of units of the goods to be shipped in order to minimize transportation expenses. Review Questions 1. Discuss the components of a transportation table. stated in Describe the three ways of obtaining the initial solution. i = £ § : E E 2 2 a Demonstrate the difference in calculating the net contributions using ‘SSM and the MODI method. Case Problems Pilipad said fuel 1. Solve the given transportation problem. areas at th Destinations Sources A B c Determine 92 MANAGEMENT SCIENCEa. Find the initial solution using NWC. b. Find the initial solution using MCM. ©. Find the initial solution using VAM. d. Find the optimal solution, 2. Pilipinas Power Corporation buys fuel every month for five Metro Manila areas. ‘The requirements for each area (in hundreds of thousands of drums) are as follows: Area Location Requirement Taguig City 55 Muntiniupa City 65 Pateros 20 Las Pifias City 44 §: Paranaque City 76 é ‘Three companies from foreign countries have submitted their bid to supply the demand for fuel: ‘ Foreign Countries Maximum Capacity USA’ 80 Russia 60 China 120 The transportation costs for each company and each Metro Manila area are stated in the table below in dollar currency: Suppliers | Taguig | Muntinlupa | Pateros | Las Pifias | Paranaque USA 18 1.6 13 06 03 Russia [1.6 12 09 02 06 China| 0.8 1.0 44 12 1.6 Pilipinas Power Company can order the entire quantity or portions from the Said fuel suppliers. The company aims to supply the demand of the five Metro Manila areas at the minimum total cost. Determine the contract award policy that would be followed by the company, CHAPTER7 DISTRIBUTION MODELS PART1 93