LINEAR PROGRAMMING:
Simplex Method
The simplex method, devised by George
B. Dantzig in 1947, is a highly efficient method
of solving more complex linear programming
problems. It starts with an initial basic feasible
solution and then proceeds step by step to
subsequent solutions each of which is a basic
feasible solution. It is an iterative process – the
computational process is repeated over and
over following a standard pattern until the best
solution is obtained (step by step process
following a standard pattern).
Sources: Quantitative Techniques by Sirug/Tabuloc
Simplex Method
• The simplex method of solving linear
programming problems is a set of
mathematical steps with the use of tables
that provide a direct and economical way of
solving LPP.
• This method is also called tabular method
since the mathematical steps used are
presented in tabular form.
A. Standard Linear Programming Maximization
A standard linear programming
maximization model are required to maximize
an objective function of the form
Maximize: P = a1x1 + a2x2 + a3x3 + … + anxn
Subject to:
b1 x1 b2 x2 b3 x3 ... bn xn c structural constraints
x1 0, x2 0, x3 0,..., xn 0 non - negativity constraints
Where c > 0 and the inequality in the structural
constraints is strictly “ ≤ ”.
B. Standard Linear Programming Minimization
A standard linear programming
minimization model are required to minimize an
objective function of the form
Minimize: C = a1x1 + a2x2 + a3x3 + … + anxn
Subject to:
b1 x1 b2 x2 b3 x3 ... bn xn c structural constraints
x1 0, x2 0, x3 0,..., xn 0 non - negativity constraints
Where c > 0 and the inequality in the structural
constraints is strictly “ ≥ ”.
Exercise:
Determine whether the following are standard LP
model or non-standard LP model:
1. Maximize : P 10 x1 22 x2 15 x3
Subject to : x1 x2 x3 355
2 x1 2 x2 x3 500
x1 4 x2 5 x3 420
x1 0, x2 0, x3 0
2. Minimize : C 12 x1 11x2 16 x3
Subject to : x1 x2 x3 370
2 x1 x2 x3 230
6 x1 8 x2 5 x3 480
x1 0, x2 0, x3 0
3. Maximize : P 3x1 2 x2 x3
Subject to : x1 x2 x3 30
2 x1 x2 x3 30
6 x1 8 x2 5 x3 30
x1 0, x2 0, x3 0
4. Minimize : C 6 x1 x2 9 x3
Subject to : x1 5 x3 65
x2 2 x3 80
3x1 5 x2 20
x1 0, x2 0, x3 0
Important Terms:
Standard Form – is expressing the constraints of a Linear
Programming Problem as equalities with all the variables on the
left side of the equation and constant on the right side.
Slack Variables – are variables added to the constraints to convert
inequality with ≤ or < into equations. The slack variables are the
unused resources such as machine time, labor hours, raw
materials, etc.
Surplus Variables – are variables subtracted to the constraints to
convert inequality with ≥ or > into equations.
Iteration – is a simplex method which consists of the sequence of
steps (row operations) performed in moving one basic feasible
solution to another.
Important Terms:
Right-Hand-Side (RHS) – is the column in a simplex tableau
indicating the quantities of the variables (basic variables) in a
solution (basic solution).
Basic Variables (BV) – are the variables included in a basic solution
which has only one non zero element along the column
corresponding to that variable. The value of the basic variable is
found along the RHS column.
Non-basic Variables (NBV) – are the variables which has more than
one non-zero element along the column corresponding to that
variable. The variable is non-basic because it doesn’t contribute
to the basic solution of the problem and the value of that variable
is zero.
Pivot column – is the column in any solution to a maximization
problem which has the most positive value in the Cj - Zj row or
minimization problem which has the most negative value in the
Cj - Zj row.
Pivot row – is the row in the simplex tableau corresponding to the
basic variable that will leave the solution. It is determined by the
test ratio and it is being computed by dividing the right-hand-side
(RHS) by the corresponding elements in the pivot column.
• The test ratio must be a nonnegative real number.
• The smallest positive ratio will determine the pivot row of the
simplex tableau.
• In cases that there are two or more equal test ratios, choose any
one.
Pivot – is the element of the simplex tableau that is in both the pivot
row and the pivot column.
Pivoting – the process of going from one simplex tableau to the next is
called pivoting. (We pivot in order to make the non-basic variable
become a basic variable.)
Standard Linear Programming Model: Maximization
Problem
How to set up the initial table
1. Transform the inequalities (explicit constraints) to equality form by
adding slack variables. In terms of constraints, we may interpret
slack variables as unused resources; they arise with “≤”
constraints.
Explicit constraints and Modified Explicit constraints
2. Include the slack variables in the non-negativity constraints.
Non-negativity Constraints and Modified Non-negativity Constraints
3. Rewrite the objective function by which includes the slack
variables whose coefficients are zero. ( zero coefficients because
slacks variable do not contribute to the Objective function).
Objective Function and Modified Objective Function
4. Set up the initial simplex table by entering the variables, the
objective function and the constraint coefficients.
The table below illustrates the simplex tableau:
1. The contribution to profit of each variable, cj, is indicated above each variable in the
first row.
2. The basic variables are the variables in the current solution. The slack variables
corresponding to each constraint are the basic variables in the initial simplex tableau.
3. The coefficients of the variables in each constraint are written in the cells
corresponding to each slack variable (basic variable).
4. The entries in the cells in the Zj row, j = x1, x2, x3, S1, S2, S3, are obtained as follows:
Z j a jc j Where: aj is the cell entry in the column corresponding to each cj.
5. The entries of the Cj -Zj row will determine if the solution is optimal The solution is
optimal if Cj - Zj ≤ 0. Otherwise, improve the solution.
OF 30 30 10 0 0 0
cj BV x1 x2 x3 S1 S2 S3
RHS/ QTY.
0 S1 2 1 2 1 0 0 14
0 S2 2 4 1 0 1 0 26
0 S3 2 4 6 0 0 1 56
ZJ 0 0 0 0 0 0
0
CI - ZJ 30 30 10 0 0 0
5. Examine the simplex tableau for optimal solution by examining the
Cj – Zj row.
a. If the entries in these row are less than or equal to zero, the
solution is optimal.
b. If there is one or more entry which is greater than zero, the
solution is optimal.
6. Establish a new simplex tableau: select the pivot column - the
column with the most positive entry in the Cj – Zj row. Find the
pivot row - the row with the smallest positive ratio. The Test Ratio
is computed by dividing the RHS value by the corresponding entry
of the basic variable (Sn) in the pivot column.
7. Identify the Entering Variable (incoming variable) and the Exiting
Variable (Leaving Variable).
• If all the entries in this column (pivot column) are negative,
the problem is unbounded and there is no solution.
8. Do the next iteration by utilizing the elementary row
operations in such a way that the pivot element will be
converted to one and the entries above and below the pivot
element corresponding to all basic variables are converted
to zero. To do the elementary operations:
1
a. New row of entering variable * old row (exiting variable)
pivot element
b. The remaining rows are obtained using the elementary
row operation:
New row = c* new row of the entering variable + old row
9. Go back to step 5. If an optimal solution is reached, state the
decision to be made.
Maximization Problem
Illustration 1.
A couturier has the following materials
available: 18 square meter sateen, 20 square
meter Lace, and 5 square meter chiffon. A gown
requires the following: 3 square meter sateen, 2
square meter lace, and 1 square meter chiffon. A
cocktail dress requires the following: 2 square
meter sateen, and 4 square meter lace. If a gown
sells for P1,200 and a cocktail dress sells P1,600,
how many of each garments should the couturier
make to obtain the maximum amount of money?
Let x = be the number of gowns
y = be the number of cocktail dress
x gowns y cocktail dresses Resources
Sateen 3 2 18
Lace 2 4 20
Chiffon 1 5
cost P1200 P1600
Objective Function: Max P = 1200x +1600y
Constraints:
3x +2y ≤ 18 (Sateen constraints)
2x +4y ≤ 20 (Lace constraints)
x ≤5 (Chiffon constraints)
x, y ≥ 0 (non-negativity constraints)
Solve by simplex maximization:
Example 2.
Dimensions by Marchelle, makes two wood
items, tables and chairs, which must be processed
through assembly and finishing departments.
Assembly has 60 hours available per week; finishing
can handle up to 48 hours of work a week.
Manufacturing one table requires 4 hours in assembly
and 2 hours in finishing. Each chair requires 2 hours in
assembly and 4 hours in finishing. If profit is P7000
per table and P4000 per chair, determine the number
of tables and chairs that must be produced to
maximize the profit.
