# Algebraic Solution of LPPs - Simplex

Method
To solve an LPP algebraically, we first put it
in the standard form. This means all
decision variables are nonnegative and all
constraints (other than the nonnegativity
restrictions) are equations with nonnegative
RHS.
Converting inequalities into equations
2 1
3 2 x x z + =
Subject to
0 ,
6 2 3
6 3
2 1
2 1
2 1
>
s +
s +
x x
x x
x x
Maximize
Consider the LPP
We make the ≤ inequalities into equations by
adding to each inequality a “slack” variable
(which is nonnegative). Thus the given LPP
can be written in the equivalent form

2 1
3 2 x x z + =
Subject to
0 , , ,
6 2 3
6 3
2 1 2 1
2 2 1
1 2 1
>
= + +
= + +
s s x x
s x x
s x x
are slack variables.
2 1
, s s
Maximize
Thus we seem to have complicated the
problem by introducing two more variables;
but then we shall see that this is easier to
solve. This is one of the “beauties” in
mathematical problem solving.
The ≥ inequalities are made into equations
by subtracting from each such inequality a
“surplus” (non-negative) variable.

Thus the LPP
2 1
3 2 x x z + =
Subject to
0 ,
2 2 3
6 3
2 1
2 1
2 1
>
> +
s +
x x
x x
x x
Maximize
is equivalent to the LPP
2 1
3 2 x x z + =
Subject to
0 , , ,
2 2 3
6 3
2 1 2 1
2 2 1
1 2 1
>
= ÷ +
= + +
s s x x
s x x
s x x
is a slack variable; is a surplus variable.
2
s
1
s
Maximize
If in a constraint, the RHS constant is
negative, we make it positive by
multiplying the constraint by -1.
Thus the LPP
2 1
3 2 x x z + =
Subject to
0 ,
2 2 3
6 3
2 1
2 1
2 1
>
÷ > +
s +
x x
x x
x x
Maximize
is equivalent to the LPP
2 1
3 2 x x z + =
Subject to
1 2
1 2
1 2
3 6
3 2 2
, 0
x x
x x
x x
+ s
÷ ÷ s
>
Maximize
Its standard form is the LPP
2 1
3 2 x x z + =
Subject to
0 , , ,
2 2 3
6 3
2 1 2 1
2 2 1
1 2 1
>
= + ÷ ÷
= + +
s s x x
s x x
s x x
Maximize
are slack variables.
2 1
, s s
Dealing with unrestricted variables
If, in an LPP, a decision variable x
i
is
unrestricted (in sign) i.e. it can take positive
as well as negative values, then we can, by
writing
i i i
x x x
+ ÷
= ÷
,
i i
x x
+ ÷
are (defined below and are) nonnegative, make
the LPP into an equivalent LPP where all the
decision variables are ≥ 0.
Note:
| |
;
2
i i
i
x x
x
+
+
=
if 0 and otherwise
i i i i
x x x x
+ ÷
= > ÷
where
| |
2
i i
i
x x
x
÷
÷
=
Thus the LPP
Maximize
2 1
3x x z + =
Subject to
1 2
1 2
1 2
2
4
unrestricted, 0
x x
x x
x x
+ s
÷ + s
>
is equivalent to the LPP
2 1 1
3x x x z + ÷ =
÷ +
Subject to
1 1 2 1
1 1 2 2
1 1 2 1 2
2
4
, , , , 0
x x x s
x x x s
x x x s s
+ ÷
+ ÷
+ ÷
÷ + + =
÷ + + + =
>
Maximize
Basic variables, Basic feasible Solutions
Consider an LPP (in standard form) with m
constraints and n decision variables. We
assume m ≤ n. We choose n –m variables and
set them equal to zero. Thus we will be left
with a system of m equations in m variables.
If this m×m square system has a unique
solution, this solution is called a basic
solution. If further if it is feasible, it is called
a Basic Feasible Solution (BFS).
The n-m variables set to zero are called
nonbasic and the m variables which we are
solving for are known as basic variables.
Thus a basic solution is of the form
x = (x
1
, x
2
, …, x
n
) where n-m “components”
are zero and the remaining m components
form the unique solution of the square
system (formed by the m constraint
equations).
Note that we may have a maximum of
basic solutions.
n
m
| |
|
\ .
Consider the LPP:
Maximize
2 1
3 2 x x z + =
Subject to
0 ,
6 2 3
6 3
2 1
2 1
2 1
>
s +
s +
x x
x x
x x
C
This is equivalent to the LPP (in standard form)
Maximize
2 1
3 2 x x z + =
Subject to
0 , , ,
6 2 3
6 3
2 1 2 1
2 2 1
1 2 1
>
= + +
= + +
s s x x
s x x
s x x
are slack variables.
2 1
, s s
Nonbasic
(zero)
variables
Basic
variables
Basic
solution
Assoc
-iated
corner
point
Feasible?
Object-
ive value,
z
(0,0,6,6) O Yes 0
(0,2,0,2) B Yes 6
(0,3,-3,0) E No -
(6,0,0, -12) D No -
(2,0,4,0) A Yes 4
C Yes
48/7
Optimal
) 0 , 0 ,
7
12
,
7
6
(
) , , , (
2 1 2 1
s s x x
) , (
2 1
x x
) , (
1 1
s x
) , (
2 1
s x
) , (
1 2
s x
) , (
2 2
s x
) , (
2 1
s s ) , (
2 1
x x
) , (
1 1
s x
) , (
2 1
s x
) , (
1 2
s x
) , (
2 2
s x
) , (
2 1
s s
Graphical solution of the above LPP
x1
x2
O A
D
B
E
C
(2,0)
(6,0)
(0,2)
(0,3)
(6/7, 12/7)
Optimal point
(0,0)
Thus every Basic Feasible Solution
corresponds to a corner(=vertex) of the set
S
F
of all feasible solutions.
S
F

Consider the LPP:
Maximize
2 1
3x x z + =
Subject to
1 2
1 2
1 2
unrestricted
2
4
, 0
x x
x x
x x
+ s
÷ + s
>
Question 6 (Problem set 3.2A – Page 79)
This is equivalent to the LPP(in standard form)
Maximize
2 1 1
3x x x z + ÷ =
÷ +
Subject to
0 , , , ,
4
2
2 1 2 1 1
2 2 1 1
1 2 1 1
>
= + + + ÷
= + + ÷
÷ +
÷ +
÷ +
s s x x x
s x x x
s x x x
Nonbasic
(zero)
variables
Basic
variables
Basic
solution
Assoc-
iated
corner
point
Feasible? Objective
value, z
(0,0,0,2,4) O Yes 0
(0,0,2,0,2) B Yes 6
(0,0,4,-2,0) E No -
(0,-2,0,0,6) - No -
(0,4,0,6,0) D Yes -4
(0,1,3,0,0) C Yes 8
) , , , , (
2 1 2 1 1
s s x x x
÷ +
) , , (
1 1 1
s x x
÷ +
) , , (
1 2 1
s x x
+
) , , (
2 1 1
x x x
÷ +
) , , (
2 1 1
s x x
÷ +
) , , (
2 2 1
s x x
+
) , , (
2 1 1
s s x
+
) , (
2 1
s s
) , (
2 1
x x
÷
) , (
1 1
s x
÷
) , (
2 1
s x
÷
) , (
1 2
s x
) , (
2 2
s x
Optimal value
Nonbasic
(zero)
variables
Basic
variables
Basic
solution
Associ-
ated
corner
point
Feasible
?
Objective
value, z
(2,0,0,0,6) A Yes 2
(-4,0,0,6,0) - No -
(-1,0,3,0,0) - No -
N0
Solution
- - -
) , , , , (
2 1 2 1 1
s s x x x
÷ +
) , , (
2 2 1
s x x
÷
) , , (
2 1 2
s s x
) , , (
2 1 1
s s x
÷
) , , (
1 2 1
s x x
÷
) , (
1 1
÷ +
x x
) , (
2 1
x x
+
) , (
1 1
s x
+
) , (
2 1
s x
+
Hence note that the number of Basic Solutions can be less
than
n
m
| |
|
\ .
(-1,3)
z maximum 8
at
0
B
E
D
A
C
Direction of
increasing z
2 1
3x x z + =
Example: Convert the following
optimization problem into a LPP:
Maximize
1 2 1 2
max{| 2 3 |, | 3 7 |} z x x x x = + ÷
Subject to
1 2
1 2
1 2
2
4
, 0
x x
x x
x x
+ s
÷ + s
>
Note that here the objective function is
NOT linear. Let us put
1 2 1 2
max{| 2 3 |, | 3 7 |} y x x x x = + ÷
Hence
1 2 1 2
| 2 3 | and | 3 7 | y x x y x x > + > ÷
Which is equivalent to
1 2 1 2
2 3 , (2 3 ) y x x y x x > + > ÷ +
( )
1 2 1 2
3 7 , 3 7 y x x y x x > ÷ > ÷ ÷
Hence the given optimization problem
is equivalent to the LPP:
Maximize z y =
Subject to
1 2
1 2
2
4
x x
x x
+ s
÷ + s
1 2 1 2
2 3 0, 2 3 0 x x y x x y + ÷ s ÷ ÷ ÷ s
1 2 1 2
1 2
3 7 0, 3 7 0.
, , 0
x x y x x y
x x y
÷ ÷ s ÷ + ÷ s
>