EMIS 8374 LP Review: The Ratio Test Main Steps of the Simplex Method 1. 2. 3. 4. 5. 6. Put the problem in row-0 form. Construct the simplex tableau. Obtain an initial BFS. If the current BFS is optimal then goto step 9. Choose a non-basic variable to enter the basis. Use the ratio test to determine which basic variable must leave the basis. 7. Perform the pivot operation on the appropriate element of the tableau. 8. Goto Step 4. 9. Step 9. Stop. 1 Example 1: Step 1 Original LP Maximize 4.5 x1 + 4 x2 s.t. 30 x1 + 12 x2 6000 10 x1 + 8 x2 2600 4 x1 + 8 x2 2000 x1, x2 0 LP in Row-0 Form Maximize z s.t. z - 4.5 x1 - 4 x2 = 0 30 x1 + 12 x2 + x3 = 6000 10 x1 + 8 x2 + x4 = 2600 4 x1 + 8 x2 + x5 = 2000 x1, x2, x3, x4, x5 0 2 Example 1: Steps 2 and 3 z x1 x2 x3 x4 x5 b 1 -4.50 -4 0 0 0 basis 0z 0 30 12 1 0 0 6000 x3 0 10 8 0 1 0 2600 x4 0 4 8 0 0 1 2000 x5 Initial BFS: BV = {z, x3, x4, x5}, NBV = {x1, x2} z = 0, x3 = 6,000, x4 = 2,600, x5 = 2,000 x1 = x2 = 0 3 Example 1: Steps 4 and 5 z x1 x2 x3 x4 x5 b 1 -4.50 -4 0 0 0 basis 0z 0 30 12 1 0 0 6000 x3 0 10 8 0 1 0 2600 x4 0 4 8 0 0 1 2000 x5 x1 and x2 are eligible to enter the basis. Select x1 to become a basic variable 4 Example 1: Step 6 • How much can we increase x1? • Constraint in Row 1: 30 x1 + 12 x2 + x3 = 6000 => x3 = 6000 - 30 x1 - 12 x2. • x2 = 0 (it will stay non-basic) • x3 0 • Thus x1 200. 5 Example 1: Step 6 • How much can we increase x1? • Constraint in Row 2: 10 x1 + 8 x2 + x4 = 2600 => x4 = 2600 - 10 x1 - 8 x2 • x2 = 0 (it will stay non-basic) • x4 0 • Thus x1 260. 6 Example 1: Step 6 • • How much can we increase x1? Constraint in Row 3: 4 x1 + 8 x2 + x5= 2000 => x5 = 2000 - 4 x1 - 8 x2 • • • x2 = 0 (it will stay non-basic) x5 0 Thus x1 500. 7 Example 1: Step 6 • From constraint 1, we see that we can increase x1 up to 200, but we must reduce x3 to zero to satisfy the constraints. • From constraint 2, we see that we can increase x1 up to 260, but we must also reduce x4 to zero to satisfy the constraints. • From constraint 3, we see that we can increase x1 up to 500, but we must reduce x5 to zero to satisfy the constraints. • Since x3 is the limiting variable, we make it nonbasic as x1 becomes basic. 8 Step : The Ratio Test Row 1: 30 x1 + 12 x2 + x3 = 6000 => 30 x1 + x3 = 6000 => x1 6000/30 = 200. Row 2: 10 x1 + 8 x2 + x4 = 2600 => 10 x1 + x4 = 2600 => x1 2600/10 = 260. Row 3: 4 x1 + 8 x2 + x5= 2000 => 4 x1 + x5 = 2000 => x1 2000/4 = 500. 9 Example 1: Ratio Test z x1 x2 x3 x4 x5 b 1 -4.50 -4 0 0 0 basis ratio 0z 0 30 12 1 0 0 6000 x3 200 0 10 8 0 1 0 2600 x4 260 0 4 8 0 0 1 2000 x5 500 The minimum ratio occurs in row 1. Thus, x3 leaves the basis when x1 enters. 10 Example 1: Steps 7 (Pivot) z x1 x2 x3 x4 x5 b 1 -4.50 -4 0 0 0 basis 0z 0 30 12 1 0 0 6000 x3 0 10 8 0 1 0 2600 x4 0 4 8 0 0 1 2000 x5 Pivot on the x1 column of row 1 to make x1 basic and x3 non-basic. 11 Example 1: Steps 7 (Pivot) First ERO: divide row 1 by 30 z x1 x2 x3 1 -4.50 0 -4 x4 x5 b 0 0 0 basis 0z 1 0.40 0.0333 0 0 200 x1 0 10 8 0 1 0 2600 x4 0 4 8 0 0 1 2000 x5 12 Example 1: Steps 7 (Pivot) Second ERO: Add –10 times row 1 to row 2 z x1 x2 x3 1 -4.50 -4 x4 x5 b 0 0 0 basis 0z 0 1 0.40 0.0333 0 0 200 x1 0 0 4 -0.333 1 0 600 x4 0 4 8 0 0 1 2000 x5 13 Example 1: Steps 7 (Pivot) Third ERO: Add –4 times row 1 to row 3 z x1 x2 x3 1 -4.50 -4 x4 x5 b 0 0 0 basis 0z 0 1 0.40 0.0333 0 0 200 x1 0 0 0 0 6.4 -0.133 0 1 1200 x5 4 -0.333 1 0 600 x4 14 Example 1: Steps 7 (Pivot) Fourth ERO: Add 4.5 times row 1 to row 0 z x1 x2 x3 x4 x5 b basis 1 0 -2.20 0 1 0.40 0.0333 0 0 200 x1 0 0 0 0 6.4 -0.133 0 1 1200 x5 0.15 0 0 900 z 4 -0.333 1 0 600 x4 15 Example 1: Steps 4 z x1 x2 x3 x4 x5 b basis 1 0 -2.20 0.15 0 0 900 z 0 1 0.40 0.0333 0 0 200 x1 0 0 0 0 6.4 -0.133 0 1 1200 x5 4 -0.333 1 0 600 x4 BV = {z, x1, x4, x5}, NBV = {x2, x3} z = 900, x1 = 200, x4 = 600, x5 = 1200 Increasing x2 may lead to an increase in z. 16 Example 1: Ratio Test z x1 x2 x3 x4 x5 b basis ratio 1 0 -2.20 0.15 0 0 900 z 0 1 0.40 0.0333 0 0 200 x1 500 0 0 4 -0.333 1 0 600 x4 150 0 0 6.4 -0.133 0 1 1200 x5 188 The minimum ratio occurs in row 2. Thus, x4 leaves the basis when x2 enters. 17 Example 1: Pivoting x2 into the basis z x1 x2 x3 x4 x5 b basis 1 0 0 0 0.4167 0.0833 1250 z 0 1 0 0 0.1667 -0.1667 100 x1 0 0 1 0 -0.0833 0.2083 200 x2 0 0 0 1 -4 2.5 600 x3 BV = {z, x1, x2, x3}, NBV = {x4, x5} z = 1250, x1 = 100, x2 = 200, x3 = 600 This an optimal BFS. 18