SIMPLEX METHOD TO SOLVE LPP

STEP1
Convert LPP into STANDARD FORM in the following manner
a) Addition of slack variable : It represents a quantity of
unused resources either in the form of time on a machine,
labour hrs, warehouse space, quantity of raw material etc.
It is added to less than (≤) type constraints in order to
convert it into equality. Coefficient of slack variable in the
objective function is 0
e.g
𝑀𝑎𝑥 𝑧 = 𝑥1 + 5𝑥2 + 4𝑥3
Subject to
. 2𝑥1 + 3𝑥2 + 7𝑥3 ≤ 30
𝑥1 + 2𝑥2 + 4𝑥3 ≤ 45
X1 ≥ 0, 𝑥2 ≥ 0, 𝑥3 ≥ 0
Each inequality is converted into equality by adding slack
variable 𝑠1 & s2 whose coefficient in the objective
function is 0
Max z =x1+5x2+4x3+0s1+0s2
Subject to
. 2𝑥1 + 3𝑥2 + 7𝑥3 + 𝑠1 = 30
𝑥1 + 2𝑥2 + 4𝑥3 + 𝑠2 = 45
𝑥1 ≥ 0, 𝑥2 ≥ 0, 𝑥3 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0
𝑥1 , 𝑥2 & 𝑥3 are known as decision variables
s1 & s2 are known as slack variables
 b) Subtraction of Surplus variable: It represents the amount
by which solution values exceeds a resources . Therefore It
is subtracted from the constraint which are in the form of
greater than (≥) type constraints in order to convert it into
equality. Coefficient of surplus variable in the objective
function is 0
𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2
Subject to
e.g. 2𝑥1 + 4𝑥2 ≥ 40
3𝑥1 + 5𝑥2 ≥ 20
x1 ≥ 0, 𝑥2 ≥ 0
each inequality is converted into equality by subtracting
surplus variable s1 & s2 whose coefficient in the objective
function is 0
 𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2
Subject to
e.g. 2𝑥1 + 4𝑥2 − 𝑠1 = 40
3𝑥1 + 5𝑥2 − 𝑠2 = 20
x1 ≥ 0, 𝑥2 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0
𝑥1 & 𝑥2 are known as decision variables
𝑠1 & 𝑠2 are known as surplus variables
c) Addition of Artificial variable: These are added to those
constraints which has surplus variables& its coefficient in
the objective function is M in case of 𝑚𝑖𝑛 𝑧 & −𝑀 in case
of 𝑚𝑎𝑥 𝑧 where M is very large positive cost
𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2
Subject to
e.g. 2𝑥1 + 4𝑥2 ≥ 40
3𝑥1 + 5𝑥2 ≥ 20
x1 ≥ 0, 𝑥2 ≥ 0
𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2
Subject to
e.g. 2𝑥1 + 4𝑥2 − 𝑠1 = 40
3𝑥1 + 5𝑥2 − 𝑠2 = 20
x1 ≥ 0, 𝑥2 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0
𝑥1 & 𝑥2 are known as decision variables
𝑠1 & 𝑠2 are known as surplus variables
𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2 + 𝑀𝐴1 + 𝑀𝐴2
Subject to
e.g. 2𝑥1 + 4𝑥2 − 𝑠1 + 𝐴1 = 40
3𝑥1 + 5𝑥2 − 𝑠2 + 𝐴2 = 20
x1 ≥ 0, 𝑥2 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0
𝑥1 & 𝑥2 are known as decision variables
𝑠1 & 𝑠2 are known as surplus variables
𝐴1 & 𝐴2 are known as artificial variables

eg
𝑀𝑎𝑥 𝑧 = 𝑥1 + 𝑥2
Subject to
e.g. 2𝑥1 + 4𝑥2 ≤ 40
3𝑥1 + 5𝑥2 ≥ 20
x1 ≥ 0, 𝑥2 ≥ 0
𝑀𝑎𝑥 𝑧 = 𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2
Subject to
e.g. 2𝑥1 + 4𝑥2 + 𝑠1 = 40
3𝑥1 + 5𝑥2 − 𝑠2 = 20
x1 ≥ 0, 𝑥2 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0
𝑥1 & 𝑥2 are known as decision variables

𝑠1 is known as slack variable

𝑠2 is known as surplus variables
𝑀𝑎𝑥 𝑧 = 𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2 − 𝑀𝐴1
Subject to
e.g. 2𝑥1 + 4𝑥2 + 𝑠1 = 40
3𝑥1 + 5𝑥2 − 𝑠2 + 𝐴1 = 20
x1 ≥ 0, 𝑥2 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0, 𝐴1 ≥ 0
𝑥1 & 𝑥2 are known as decision variables

𝑠1 is known as slack variable

𝑠2 is known as surplus variables
A1 is known as artificial variable
eg
𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2
Subject to
e.g. 2𝑥1 + 4𝑥2 ≥ 40
3𝑥1 + 5𝑥2 ≤ 20
x1 ≥ 0, 𝑥2 ≥ 0

𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2

Subject to
e.g. 2𝑥1 + 4𝑥2 − 𝑠1 = 40
3𝑥1 + 5𝑥2 + 𝑠2 = 20
x1 ≥ 0, 𝑥2 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0
𝑥1 & 𝑥2 are known as decision variables

𝑠1 is known as surplus variable

𝑠2 is known as slack variables

𝑀𝑖𝑛 𝑧 = 𝑥1 + 𝑥2 + 0𝑠1 + 0𝑠2 + 𝑀𝐴1

Subject to
e.g. 2𝑥1 + 4𝑥2 − 𝑠1 + 𝐴1 = 40
3𝑥1 + 5𝑥2 + 𝑠2 = 20
x1 ≥ 0, 𝑥2 ≥ 0, 𝑠1 ≥ 0, 𝑠2 ≥ 0, 𝐴1 ≥ 0
𝑥1 & 𝑥2 are known as decision variables

𝑠1 is known as surplus variable

𝑠2 is known as slack variables
A1 is artificial variable
Solve the following LPP by Simplex Method
i) 𝑀𝑎𝑥 𝑧 = 4𝑥 + 7𝑦
𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 𝑥 + 3𝑦 ≤ 18
4𝑥 + 2𝑦 ≤ 8
𝑥 ≥ 0, 𝑦 ≥ 0
STEP 1 To Convert given LPP into standard Form
𝑀𝑎𝑥 𝑧 = 4𝑥 + 7𝑦 + 0𝑆1 + 0𝑆2
𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 𝑥 + 3𝑦 + 𝑆1 + 0𝑆2 = 18
4𝑥 + 2𝑦 + 0𝑆1 + 𝑆2 = 8
𝑥 ≥ 0, 𝑦 ≥ 0, 𝑆1 ≥ 01 , 𝑆2 ≥ 0
𝑥, 𝑦 are decision variables & 𝑆1 , 𝑆2 are slack variable

STEP 2 EXPRESS THE ABOVE STANDARD FORM INTO MATRIX

FORM 𝑨𝑿 = 𝑩
𝑥
1 3 1 0 𝑦 18
𝐴= , 𝑋= 𝑆1 , 𝐵 =
4 2 0 1 8
𝑆2
, 𝐶𝑗 = 4 7 0 0 Cj:Matrix of Coefficient of each
variable in the objective function
STEP 3 TO FIND INITIAL SOLUTION BY SUBSTITUTING 𝑥 = 0 & 𝑦 = 0 in
standard form

∴ initial solution is 𝑆1 = 18 , 𝑆2 = 8

STEP 4 FORM THE INITIAL SIMPLEX TABLE

STEP 5 FIND THE VALUES OF 𝒁𝒋 = (𝑪𝑩 𝒙𝒋 ) for each variable

STEP 6 FIND ∆𝒋 = 𝑪𝒋 − 𝒁𝒋

STEP 7 FIND INCOMING VARIABLE CORRESPONDING TO HIGHEST ∆𝒋

STEP 8 FIND REPLACEMENT RATIO 𝑋𝐵 /𝑥𝑗

STEP 9 FIND OUTGOING VARIABLE CORRESPONDING TO MINIMUM

POSITIVE RATIO

STEP 10 FIND KEY ELEMENT i.e. intersecting element of key column &
key row

STEP 11 MODIFY THE VALUES by performing calculations

Old element-(New element * the element which is to be converted into

zero)

& REPEAT STEPS 5, 6, …. UNTILL WE GET ALL VALUES OF ∆𝒋 ≥ 𝟎

Simplex Table
Basic 𝑪𝒋 4 7 0 0 Replacement
Variables ratio=(𝑿𝑩 /𝒙𝒋)
𝑿𝑩 𝑪𝑩 𝒙 𝒚 𝑺𝟏 𝑺𝟐

KEY COLUMN
S1(OLD) 18 0 1 3 1 0 18/3=6

S2 0 4 2 Key 0 1 8/2=4
(KEY 8 element (Outgoing vari
ROW)
0×1+0 0×3+0 0×1+0 0×0
𝒁𝒋 ×4 ×2 =0+0 ×0 +0
=0+0 =0 =0+0 ×1
= (𝑪𝑩 𝒙𝒋 )
=0 =0 =0+0
=0
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 4−0=4 7−0=7 0−0= 0 0−0
Incoming =0
Variable

18 0 1 − (2 ∗ 3) 3 − (1 ∗ 3) 1 − (0 ∗ 3) 0
S1 − (4 =1−6 =3−3=0 =1−0 − (1/2
∗ 3) = −5 =1 ∗ 3)
= 18 =0
− 12 − 3/2
=6 = 3/2
(NEW)*3 8/2=4 7 4/2=2 2/2=1 0/2=0 ½
y
𝒁𝒋 0*(- 0*0+7*1=7 0*1+7*0=0 0*3/2+7
5)+7*2=0+ *1/2=7/
= (𝑪𝑩 𝒙𝒋 )
14=14 2
4 − 14 7-7=0 0-0=0 0 − 7/2 = −7/2
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 = −10
Since there is no ∆𝒋 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 solution so obtained is optimal
Solution of given LPP is x=0 ,y=4 zmax=28

ii) 𝑀𝑎𝑥 𝑧 = 50𝑥 + 20𝑦

𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 20𝑥 + 10𝑦 ≤ 500 , 50𝑥 + 50𝑦 ≤ 300 ,
𝑥 ≥ 0, 𝑦 ≥ 0
STEP I STANDARD FORM
𝑀𝑎𝑥 𝑧 = 50𝑥 + 20𝑦 + 0𝑆1 + 0𝑆2
𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 20𝑥 + 10𝑦 + 𝑆1 = 500
50𝑥 + 50𝑦 + 𝑆2 = 300
𝑥 ≥ 0, 𝑦 ≥ 0, 𝑆1 ≥ 0, 𝑆2 ≥ 0
𝑥, 𝑦 are decision variables & 𝑆1 , 𝑆2 are slack variables

STEP 2 MATRIX FORM 𝑨𝑿 = 𝑩

𝑥
20 10 1 0 𝑦 500
𝐴= , 𝑋= 𝑆1 ,𝐵 = ,
50 50 0 1 300
𝑆2
𝐶𝑗 = 50 20 0 0
STEP 3 INITIAL SOLUTION BY SUBSTITUTING 𝑥 = 0 & 𝑦 = 0 in standard
form
∴ initial solution is 𝑆1 = 500 , 𝑆2 = 300
Simplex Table

Basic 𝑪𝒋 50 20 0 0 Replacemen
Variables
𝑿𝑩 𝑪𝑩 𝒙 𝒚 𝑺𝟏 𝑺𝟐
𝑺𝟏 500 0 20 10 1 0 500/20=25
Old
𝑺𝟐 300 0 50 0 1 6
50
Key
element 300/50=
𝒁𝒋 = (𝑪𝑩 𝒙𝒋) (0×20) + (0×10) + (0×1)+(0×0) = (0×0) +
(0×50) = (0×50) = 0 0 (0×1) = 0
0
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 50−0= 20 − 0 = 0−0= 0 0−0=0
20
50

𝑺𝟏 500−6×20 = 0 20−(1×20 10−(1×20 1−(0×20) = 1 0−(1/50×20

Old- 380 ) =0 ) = −10 ) = −2/5
new*20
𝒙 300/50=6 50 50/50=1 50/50=1 0/50=0 1/50
New
𝒁𝒋 = (𝑪𝑩 𝒙𝒋) (0× 0) + (0× −10) + (0×1) + (0× −2/5) +
(50 × 1) = (50×1) = 50 (50×0) = 0 (50×1/50) =
50 1
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 50 − 50 20 − 50 = 0−0= 0 0 − 1 = −1
=0 −30

Since there is no positive value of ∆𝒋 ,solution is optima Therefore

,optimal solution is 𝒙 = 𝟔, 𝒚 = 𝟎, 𝒛𝒎𝒂𝒙 = 𝟔 × 𝟓𝟎 + 𝟐𝟎 × 𝟎 = 𝟑𝟎

iii) 𝑀𝑎𝑥 𝑧 = 8𝑥 + 20𝑦

𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 2𝑥 + 𝑦 ≤ 80 , 3𝑥 + 4𝑦 ≤ 96 , 𝑥 ≥ 0, 𝑦 ≥ 0
STEP I STANDARD FORM
𝑀𝑎𝑥 𝑧 = 8𝑥 + 20𝑦 + 0𝑆1 + 0𝑆2
𝑆𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 2𝑥 + 𝑦 + 𝑆1 = 80
3𝑥 + 4𝑦 + 𝑆2 = 96
𝑥 ≥ 0, 𝑦 ≥ 0, 𝑆1 ≥ 0, 𝑆2 ≥ 0
𝑥, 𝑦 are decision variables & 𝑆1 , 𝑆2 are slack variables
STEP 2 MATRIX FORM 𝑨𝑿 = 𝑩
𝑥
2 1 1 0 𝑦 80
𝐴= , 𝑋= 𝑆 ,𝐵 = , 𝐶𝑗 = 8 20 0 0
3 4 0 1 1 96
𝑆2
STEP 3 INITIAL SOLUTION BY SUBSTITUTING 𝑥 = 0 & 𝑦 = 0 in standard form
∴ initial solution is 𝑆1 = 80 , 𝑆2 = 96
Simplex Table

Basic 𝑪𝒋 8 20 0 0 Replacement ra
Variables
𝑿𝑩 𝑪𝑩 𝒙 𝒚 𝑺𝟏 𝑺𝟐
𝑺𝟏 80 0 2 1 1 0 80/1=80
𝑺𝟐 96 0 3 4 0 1 96/4=
24

𝒁𝒋 (0×2) + (0×1) + (0×1)+(0× (0×0

(0×3) = 0 (0×4) = 0 0) = 0 )+
(0×1
)=0
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 8−0=8 20 − 0−0=0 0−0
20 0= =0

𝑺𝟏 80−(24× 0 2−(3/4× 1) = 1−(1×1) = 0 1−(0×1 0−(1/4

1) = 56 5/4 ) =1 ×1) =
−1/4
𝒚 96/4=24 20 3/4 4/4=1 0/4=0 1/4
𝒁𝒋 (0× 5/4) + (0× 0) + (0×1) + (0×
(20 × 3/4) = (20×1) = 20 (20×0) = 0 −1/
15 4) +
(20×1/
4) = 5
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 8 − 15 = −7 20 − 20 = 0 0−0=0 0−5
= −5
Since there is no positive value of ∆𝒋 ,solution is optimal

Therefore ,optimal solution is 𝒙 = 𝟎 , 𝒚 = 𝟐𝟒 𝒛𝒎𝒂𝒙 = 𝟖 × 𝟎 + 𝟐𝟎 × 𝟐𝟒 = 𝟒𝟖𝟎

CASES OF SOLUTION OF SIMPLEX METHOD

a) In the final simplex table of given LPP when one of

MULTIPLE SOLUTION:
the non-basic variable has ∆𝒋 = 𝒄𝒋 − 𝒛𝒋 = 𝟎 then it has Multiple
solution.
Eg
Basic Cj 6 4 0 0 0 Replacement
variables Xb Cb X1 X2 S1 S2 S3 ratio
S1 14 0 0 5/3 1 -2/3 0 42/5
S3 5 0 0 -1/3 0 1/3 1 ---
X1 8 6 1 2/3 0 1/3 0 12
zj 6 4 0 2 0
Cj-zj 0 0 0 -2 0
One of the non-basic variable x2 has cj-zj=0 therefore we
conclude that given LPP has multiple solution
b) DEGENERATE SOLUTION:

Degeneracy arises when there is a tie in the minimum

replacement ratio that determines the variable to leave the basis.
e.g.
Basic Cj 3 9 0 0 Replacement
variable Xb Cb X1 X2 S1 S2 ratio(Xb/x2)
S1 8 0 1 4 1 0 8/4=2
S2 4 0 1 2 0 1 4/2=2
Zj 0 0 0 0
Cj-zj 3 9 0 0
From the above table we observe that 𝑥2 is incoming variable. But
minimum replacement ratio is same for both s1 & s2 for outgoing
variable. i.e there is a tie among two variables . This is an indication of
existence of degeneracy.
 c) Unique Solution: If cj-zj <0 for each non-basic variable in the final
simplex table then we conclude that given LPP has unique solution
Final Simplex Table
Basic 𝑪𝒋 8 20 0 0 Replacement ra
Variables
𝑿𝑩 𝑪𝑩 𝒙 𝒚 𝑺𝟏 𝑺𝟐
𝑺𝟏 56 0 5/4 0 1 −1/4
𝒚 24 20 3/4 1 0 1/4
𝒁𝒋 15 20 0 5
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 −7 0 0 −5

Non basic variables x & s2 has cj-zj<0 therefore we conclude that

given LPP has unique solution

d) UNBOUNDED SOLUTION: If there is no positive replacement ratio

then we conclude that given LPP has unbounded solution
Basic 𝑪𝒋 60 20 0 0 Replacement ra
Variables
𝑿𝑩 𝑪𝑩 𝒙 𝒚 𝑺𝟏 𝑺𝟐
𝑺𝟐 240 0 0 10 −4 1 ---
𝒙 60 60 1 2 -1/2 0 ----
𝒁𝒋 60 120 30 0
∆𝒋 = 𝑪𝒋 − 𝒁𝒋 0 −100 30 0

Since it is not possible to find outgoing variable due to negative

replacement ratio, we conclude that given LPP has unbounded
solution

2
0