Name:

T HE S IMPLEX M ETHOD : STANDARD M AXIMIZATION P ROBLEMS

□ A linear programming problem consists of a linear objective function to be maximized or minimized
subject to certain constraints in the form of linear equations or inequalities.

□ The Simplex Method is a method of finding the corner points for a linear programming problem with n
variables algebraically.

□ STANDARD MAXIMIZATION PROBLEMS meet the following conditions:

1. The objective function is maximized
2. All variables in the problem are non-negative.
3. Each constraint can be written so that the expression containing the variables is less than or equal to a
non-negative constant.

□ Here is the SIMPLEX METHOD:

1. Set up the initial simplex tableau:
(a) Create slack variables.
(b) Rewrite the objective function so that
□ the coefficient of P is 1.

□ all variables and P are on the same side of the equal sign.

(c) Place the constraints and the objective function in the initial simplex tableau.
2. Determine whether or not the optimal solution has been reached:
(a) The optimal solution has been reached if all entries in the last row to the left of the vertical line
are non-negative.
(b) If an optimal solution has been reached, skip to step 4.
(c) If an optimal solution has not been reached, go to step 3.
3. Perform pivot operations:
(a) Locate pivot element:
□ pivot column:

(a) Are there negatives in the constants column (above the horizontal line)? If no, skip to b. If
yes, pick any negative in that row. The column for that entry is the pivot column.
(b) The column with the most negative entry in the last row to the left of the vertical line.
□ pivot row: Divide each entry in the pivot column into the corresponding entry in the constants

column. The pivot row is the row with the smallest NON-NEGATIVE such ratio. (Cannot
divide by 0)
□ pivot element: The element in both the pivot column and pivot row.

We will use an online Simplex calculator for these parts:

(b) Convert pivot element to 1 by dividing all elements in the pivot row by the pivot element.
(c) Use row operations to convert the pivot column into a unit column. (add multiples of pivot row to
other rows as needed).
(d) Return to Step 2.
4. Determine the solution:
(a) The value of the variable heading each unit column is given by the entry lying in the column of
constants in the row containing the 1.
(b) Variables heading columns not in unit form are assigned the value of 0.

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 1
permission.
 Math 1324 Section 6.4 Continued
1. Solve the linear programming problem using the Simplex Method.

Maximize P = −x + y + 2z ← Objective Function

Subject to: 3x + 4y − 3z ≤ 23 #
5x − 4y − 3z ≤ 10 Constraints (these get slack variables)
7x + 4y + 11z ≤ 30
x ≥ 0#
y≥0 ← These say these variables are non-negative
z≥0
(a) Is this a standard maximization problem?
i. Is the objective function maximized?
ii. Are all variables in the problem non-negative? Translation: are these inequalities present:
x ≥ 0, y ≥ 0, z ≥ 0 for every variable in the problem?
iii. Can each constraint be written so that the expression containing the variables is less than or
equal to a non-negative constant? Translation: Do all inequalities look like:
(sum/difference of variable terms) ≤ 0 or (sum/difference of variable terms) ≤ positive number
iv. We see these three criteria are satisfied so, yes! We have a standard maximization problem.
(b) Now to adjust the constraints so they can also go in the tableau: Now we create slack variables.
We introduce a new variable to "take up the slack" so we can change our inequalities to
equations.
i. Let’s just talk through how this works. Looking at the first constraint, 3x + 4y − 3z ≤ 23,
we see we do have an inequality. To demonstrate how to change ≤ to =, let’s pull some
numbers out of the air for demonstration purposes.
ii. Suppose x = 3, y = 4, and z = 5. If we plug in those values for each variable, we get
3(3) + 4(4) − 3(5) ≤ 23.
9 + 16 − 15 ≤ 23
10 ≤ 23
We see this is a true statement.
iii. We will replace ≤ with +s1 =. So what number do we need to add to the left-hand side so that
we have a true equation?
10 23 ≤
In this case, we need to add 13:
10 23 +13 =
iv. Let’s generalize this. We could have chosen many different values for x, y, z, and each would
have given a different number needed to pick up the slack. As we change the values for x, y,
z, the number needed changes too. We can also say it varies. We can call it the slack variable
(able to vary) and represent it in general just like we did variables x, y, z.
v. Thus, 3x + 4y − 3z ≤ 23 becomes 3x + 4y − 3z +s1 = 23
vi. Constraint 1 becomes: 3x + 4y − 3z +s1 = 23
vii. Constraint 2 becomes: 5x − 4y − 3z +s2 = 10
viii. Constraint 3 becomes: 7x + 4y + 11z +s3 = 30
(c) Putting everything in the initial simplex tableau:
Equations with slack variables that came from the constraints go first. The objective function goes
last.
i. Rearrange the objective function so that the coefficient of P is 1 all variables and P are on the
same side of the equal sign. Translation: Move all terms on the right-hand side to the left-
hand side.

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 2
permission.
 ii. P = −x + y + 2z becomes x − y − 2z + P = 0
Maximize P = −x + y + Put this at the bottom
Subject to: 2z # →→→→→→ 3x + 4y − 3z +s1 = 23
3x + 4y − 3z ≤ 23 These become → → 5x − 4y − 3z +s2 = 10
5x − 4y −
7 ≤ 1030 →→→→→→ 7x + 4y + +s3 = 30
x + 4y +
3z # 11z
11z x ≥ 0 x − y − 2z + P =
0
y≥0 ← Not in simplex tableau
z≥0
iii. Now take these new equations and put them in a matrix, which we call our Initial

Simplex Tableau
x y z s1 s2 s3 P constant

3 4 −3 1 0 0 0 23

5 −4 −3 0 1 0 0 10

7 4 11 0 0 1 0 30

1 −1 −2 0 0 0 1 0
(d) Now let’s find our pivot element!
i. Look at the bottom row of the initial simplex tableau, to the left of the vertical line. Put an
arrow under the most negative entry. Hint: Look at all numbers with a negative sign. If they
were all positive, which is the largest?
ii. Did you put an arrow under the negative 2 in the z column? If so, you are correct! That means
the z column is the pivot column.
Initial Simplex Tableau
x y z s1 s2 s3 P constant

3 4 −3 1 0 0 0 23

5 −4 −3 0 1 0 0 10

7 4 11 0 0 1 0 30

1 −1 −2 0 0 0 1 0
↑
PC
constant entry
We will look at the z column and the constant column. We will divide like this: to
PC entry
find the smallest non-negative ratio.

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 3
permission.
 Initial Simplex Tableau
x y z s1 s2 s3 P constant Ratio
23
3 4 −3 1 0 0 0 23 = −# z z negative ratio so throw this one
out
−3
10
5 −4 −3 0 1 0 0 10 = −# z z negative ratio so throw this one
out
−3
30
7 4 11 0 0 1 0 30 ← PR (smallest non-negative ratio)
11

1 −1 −2 0 0 0 1 0
↑
PC
Initial Simplex Tableau
x y z s1 s2 s3 P constant

3 4 −3 1 0 0 0 23

5 −4 −3 0 1 0 0 10

7 4 11 0 0 1 0 30 ← PR

1 −1 −2 0 0 0 1 0
↑
PC
Initial Simplex Tableau
x y z s1 s2 s3 P constant

3 4 −3 1 0 0 0 23

5 −4 −3 0 1 0 0 10

7 4 11 0 0 1 0 30

1 −1 −2 0 0 0 1 0

2. Now we pivot on the circled element. Remember what pivot means? This makes a unit column, with
the circled entry 1 and all other entries in that column 0. We will use this calculator:
http://simplex.tode.cz/en/

Here is the next tableau after making that unit column:

x y z s1 s2 s3 P constant

56 3 343
54
11 11 0 1 0 11 0 11

200
76
11
−32
11 0 0 1 3
11 0 11 Successful pivot on 3, 3 entry (made unit column).
30
4 1
11
7 11 1 0 0 11 0 11

−3 2 60
11
25 11 0 0 0 11 1 11

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 4
permission.
 3. Has an optimal solution been reached?

No. There are still negatives in the bottom

row, to the left of the vertical line. Thus, we x y z s1 s2 s3 P constant
find the next pivot element. Do this in the 343
56 3
tableau given to the right. 54
11 11 0 1 0 11 0 11

(a) PIVOT COLUMN: Put an arrow under 200

−32 3
76
11 11 0 0 1 11 0 11
the most negative entry in the bottom
row, to the left of the vertical line. 30
4 1
11
7 11 1 0 0 11 0 11

(b) PIVOT ROW: Put an arrow next −3 2 60

to the smallest non-negative ratio 11
25 11 0 0 0 11 1 11
constant entry
.
PC entry
(c) Circle PIVOT ELEMENT

(d) Now check your answer below.

x y z s1 s2 s3 P constant Ratio

54 56
34
11 11 0 1 0 3
3
343
← PR
5
0 1
11 6
76
1
20 z z
−32
11 11 0 0 1 1
3 0
11 15
0 1
2
7 4
30
11 11
11
1 0 0 1
1
25 −3
11 11 0 1 60
11
↑
PC 0 0 0 1
2

1 1

4. Pivot on the circled element. We will use this calculator:

http://simplex.tode.cz/en/ Here is the next tableau after making that unit column:
x y z s1 s2 s3 P constant

11 3 49
27
28 1 0 56 0 56 0 8

68
7 0 0 4
7 1 3
7 0 36

1
13 −1 1
−22 0 1 14 0 14 0 2

3 11 57
88
281 0 0 56 0 56 1 8

5. Has an optimal solution been reached?

Yes! no more negatives in the bottom row, to the left of the vertical line. This tableau is in final form!!

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 5
permission.
 6. Draw a line through the columns that are not unit columns. Those variables are assigned a value of 0.
x y z s1 s2 s3 P constant

27 11 3 49 Thus, we know:
28 1 0 56 0 56 0 8 x=0
68 4 3
7 0 0 7 1 7 0 36 s1 = 0
−13 −1 1 1
22 0 1 14 0 14 0 2 s3 = 0

281 3 11 57
88 0 0 56 0 56 1 8

Pretend the non-unit columns are not there (like Now look at the columns remaining in the table.
this):
Read off the answers like we did with matrices:
y z s2 P constant
49
49
y=
1 0 0 0 8 8
s2 = 36
0 0 1 0 36 1
z=
2
1
0 1 0 0 2 57
P=
8
0 0 0 1 57 8
Here are all the variables:
49 1 57
x=0 y= z= s1 = 0 s2 = 36 s3 = 0 P =
8 2 8

Now that you’ve gotten the hang of it, let’s try some more!

1. Put the linear programming problem into an initial simplex

tableau: Maximize P = 3x + 8y
Subject to: x + 5y ≤ 48
2x + 10y ≤ 65
x≥0
y≥0

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 6
permission.
 2. Put the linear programming problem into an initial simplex
tableau: Maximize P = 8x + 3y + 4z
Subject to: 2x + 4y + 3z ≤ 14
2x − y + 5z ≤ 50
4x + 9y + 11z ≤
81 x ≥ 0
y≥0
z≥0

3. Put the linear programming problem into an initial simplex

tableau: Maximize P = 15x − 23y + 35z

Subject to: 5x + 8y + 7z ≤ 154
12x + 10y + 13z ≤ 210
11x + 14y + 26z ≤ 611
0 ≤ x ≤ 200 →→→→ HINT: Rewrite as 0 ≤ x and x ≤ 200
y≥0
z≥0

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 7
permission.
 4. Given the following simplex tableau, select the next pivot element.
x y z s1 s2 P constant
3 3 −2 1 0 0 100
5 5 3 0 1 0 150
−2 −3 4 0 0 1 0

5. Given the following simplex tableau, select the next pivot element.
x y z s1 s2 s3 P constant
4 8 9 1 0 0 0 260
−4 −6 9 0 1 0 0 80
2 1 4 0 0 1 0 390
−18 −13 −3 0 0 0 1 0

6. Given the final simplex tableau, read off the solutions:

x y z s1 s2 s3 P constant
0 1 5 4 −1 0 0 1
1 0 6 3 1 0 0 2
0 0 9 8 1 1 0 2
0 0 12 0 3 0 1 6
x=
y=
z=
s1
= s2
= s3
=P
=

7. Given the final simplex tableau, read off the solutions:

x y z s1 s2 s3 P constant

1 0 3 0 1 1 0 45
5 5

0 1 — 14 0 —3 4 0 26
5 5

0 0 — 19 1 —3 2 0 11
11 2
0 0 5 0 5 0 1 600
x=
y=
z=
s1
= s2
= s3
©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 8
permission.
 =P
=

©Melanie Yosko This page may not be scanned, copied, or redistributed without Page 9
permission.