SIMPLEX METHOD

PROBLEM WITH MIXED CONSTRAINTS

SPECIAL CASES:
INFEASIBILITY
UNBOUNDED
DEGENERACY
DIFFERENT INEQUALITY CONSTRAINTS
PROBLEM WITH MIXED COSTRAINTS
 PROBLEM WITH MIXED COSTRAINTS
MAXIMISE: Z = 2X1 + 4X2
SUBJECT TO
2X1 + X2 ≤ 18
3X1 + 2X2 ≥ 30
X1 + 2X2 = 26
X1 ; X2 ≥ 0

• MAXIMISATION CASE BUT ALL CONSTRAINTS ARE NOT OF THE ‘LESS THAN TYPE’

• CONSTRAINT 1 WILL BE STANDARDISED WITH A SLACK VARIABLE

• CONSTRAINT 1 WILL BE STANDARDISED WITH A SURPLUS VARIABLE AND

ARTIFICIAL VARIABLE

• CONSTRAINT 1 WILL BE STANDARDISED WITH A SLACK VARIABLE

• SINCE THE OBJECTIVE FUNCTION IS OF THE MAX TYPE THE ARTIFICIAL VARIABLES
WILL HAVE THE NEGATIVE COEFFICIENT i.e. -M
 STANDARDISED PROBLEM

MAXIMISE: Z = 2X1 + 4X2 + 0S1 + 0S2 - MA1 - MA2

SUBJECT TO 2X1 + X2 + S1 = 18
3X1 + 2X2 - S2 + A1 = 30
X1 + 2X2 + A2 = 26
X1 ; X 2 ; S1 ; S2 ; A1 ; A 2 ≥ 0
 INITIAL TABLE
BASIS X1 X2 S1 S2 A1 A2 bi bi/X2

S1 0 2 1 1 0 0 0 18 18
A1 -M 3 2 0 -1 1 0 30 15
A2 -M 1 2** 0 0 0 1 26 13
Cj 2 4 0 0 -M -M
Zj -4M -4M 0 M -M -M
Δj = Cj - Zj 2+4M 4+4M 0 -M 0 0
 TABLE 2
BASIS X1 X2 S1 S2 A1 bi bi/X1

S1 0 3/2 0 1 0 0 5 10/3
A1 -M 2** 0 0 -1 1 4 2
X2 4 1/2 1 0 0 0 13 26
Cj 2 4 0 0 -M
Zj 2- 2M 4 0 M -M
Δj = Cj - Zj 2M 0 0 -M 0
 TABLE 3
• TABLE 3 SHOWS OPTIMAL BASIS X1 X2 S1 S2 bi
SOLUTION
S1 0 0 0 1 ¾ 2
• AS ALL Δj ARE EITHER
ZERO OR NEGATIVE X1 2 1 0 0 -1/2 2
X2 4 0 1 0 1/4 12
• NOTE THAT S2 IS A BASIC Cj 2 4 0 0
VARIABLE BUT HAS A
VALUE EQUAL TO ZERO Zj 2 4 0 0
Δj = Cj - Zj 0 0 0 0
• OPTIMAL SOLUTION IS:
X1 = 2; X2 = 12; S1 = 2

• OPTIMAL VALUE OF
OBJECTIVE FUNCTION:
Z = 2(2) + 4(12) = 52
NO COMMON FEASIBLE REGION FOR ALL THE CONSTRAINTS

INFEASIBILITY
 IDENTIFYING INFEASIBILITY
• NO FEASIBLE SOLUTION
• IN TERMS OF THE SIMPLEX METHOD, IF IN THE
FINAL SOLUTION AN ARTIFICIAL VARIABLE IS IN
THE BASIS WITH A POSITIVE VALUE THERE IS NO
FEASIBLE SOLUTION
• EXAMPLE:
MAXIMISE: Z = 20X1 + 30X2
SUBJECT TO 2X1 + X2 ≤ 40
4X1 - X2 ≤ 20
X1 ≥ 30
X1 ; X 2 ≥ 0
 STANDARDISED PROBLEM

MAXIMISE: Z = 20X1 + 30X2 + 0S1 + 0S2 + 0S3 - MA1

SUBJECT TO 2X1 + X2 + S1 = 40
4X1 - X2 + S2 = 20
X1 – S3 + A1 = 30
X1 ; X 2 ; S1 ; S2 ; S3 ; A 1 ≥ 0
 INITIAL TABLE
BASIS X1 X2 S1 S2 S3 A1 bi

S1 0 2 1 1 0 0 0 40
S2 0 4 -1 0 1 0 0 20
A1 -M 1 0 0 0 -1 1 30
Cj 20 30 0 0 0 -M
Zj
Δj = Cj - Zj
 TABLE 3
BASIS X1 X2 S1 S2 S3 A1 bi

X2 30 0 1 2/3 - 1/3 0 0 20
X1 20 1 0 1/6 1/6 0 0 10
A1 -M 0 0 - 1/6 - 1/6 -1 1 20
Cj 20 30 0 0 0 -M
Zj 20 30 70/3 + M/6 - 20/3 + M/6 M -M
Δj = Cj - Zj 0 0 -70/3 – 20/3 – M/6 -M 0
M/6

• TABLE 3 SHOWS THERE IS NO FURTHER IMPROVEMENT POSSIBLE

IN THE ABOVE SOLUTION

• AS ALL Δj ARE EITHER ZERO OR NEGATIVE

• SINCE A1 IS STILL IN THE BASIS, THE SOLUTION IS INFEASIBLE

THE OBJECTIVE FUNCTION VALUE CAN BE OPTIMISED
(INCREASED OR DECREASED) INFINITELY

UNBOUNDED
 IDENTIFYING UNBOUDEDNESS
• THE OBJECTIVE FUNCTION VALUE CAN BE
OPTIMISED (INCREASED OR DECREASED)
WITHOUT ANY LIMIT
• IN TERMS OF THE SIMPLEX METHOD, IF IN THE
FINAL SOLUTION THERE ARE NO NON-NEGATIVE
REPLACEMENT RATIOS (i.e. ALL ARE NEGATIVE)
OR THEY ARE EQUAL TO INFINITY (WHERE THE
DENOMINATOR IS ZERO)
• EXAMPLE:
MAXIMISE: Z = 10X1 + 20X2
SUBJECT TO 2X1 + 4X2 ≥ 16
4X1 + X2 ≥ 15
X1 ; X 2 ≥ 0
 STANDARDISED PROBLEM

MAXIMISE: Z = 10X1 + 20X2 + 0S1 + 0S2 - MA1 – MA2

SUBJECT TO 2X1 + 4X2 - S1 + A1= 16

4X1 + X2 - S2 + A2= 15
X1 ; X2 ; S1 ; S2 ; A1; A2 ≥ 0
 INITIAL TABLE
BASIS X1 X2 S1 S2 A1 A2 bi

A1 -M 2 4 -1 0 1 0 16
A2 -M 4 1 0 -1 0 1 15
Cj 10 20 0 0 -M -M
Zj
Δj = Cj - Zj
 TABLE 4
BASIS X1 X2 S1 S2 A1 A2 bi bi/S2

X1 10 1 5 0 -1 0 1 15 -15
S1 0 0 6 1 -2 -1 2 14 -7
Cj 10 20 0 0 -M -M
Zj 10 50 0 -10 0 10
Δj = Cj - Zj 0 -30 0 10 -M -M-10

• TABLE 4 DOES NOT SHOW THE OPTIMAL SOLUTION, AS S2 HAS A POSITIVE Δj VALUE (10)
• CALCULATING REPLACEMENT RATIOS AS bi/S2
• AS ALL REPLACEMENT RATIOS ARE EITHER NEGATIVE OR INFINITY, THE SIMPLEX
ALGORITHM TERMINATES
• THE SOLUTION IS UNBOUNDED
WHERE A BASIC VARIABLE ASSUMES A VALUE EQUAL TO
ZERO

DEGENERACY
 IDENTIFYING DEGENERACY
• EACH LPP HAS n VARIABLES AND m CONSTRAINTS
• HENCE, THERE WOULD BE m BASIC AND n-m NON-
BASIC VARIABLES AND BASIC VARIABLES ASSSUME
POSITIVE VALUES
• IN TERMS OF THE SIMPLEX METHOD, IF AT ANY POINT
A BASIC VARIABLE ASSUMES A VALUE EQUAL TO ZERO,
THE SOLUTION IS SAID TO BE DEGENERATE
• EXAMPLE:
MAXIMISE: Z = 28X1 + 30X2
SUBJECT TO 6X1 + 3X2 ≤ 18
3X1 + X2 ≤ 8
4X1 + 5X2 ≤ 30

X1 ; X2 ≥ 0
 STANDARDISED PROBLEM

MAXIMISE: Z = 28X1 + 30X2 + 0S1 + 0S2 + 0S3

SUBJECT TO 6X1 + 3X2 + S1 = 18

3X1 + X2 + S2 = 8
4X1 + 5X2 + S3 = 30
X1 ; X2 ; S1 ; S2 ; S3≥ 0
 INITIAL TABLE
BASIS X1 X2 S1 S2 S3 bi bi/X2

S1 0 6 3 1 0 0 18 6 THERE IS A TIE
AS TWO
S2 0 3 1 0 1 0 8 8 REPLACEMENT
S3 0 4 5 0 0 1 30 6 RATIOS ARE
Cj 28 30 0 0 0 EQUAL
Zj 0 0 0 0 0
Δj = Cj - Zj 28 30 0 0 0

IN THE ABOVE CASE OF TIE BETWEEN 2 REPLACEMENT RATIOS, WE CAN CHOOSE

EITHER OF THE TWO VARIABLES (S1 OR S3) AS THE OUTGOING VARIABLE

EACH TIME A TIE BETWEEN 2 REPLACEMENT RATIOS OCCURS, THE FOLLOWING

TABLE WILL SHOW A DEGENERATE SOLUTION
WITH S1 AS THE OUTGOING VARIABLE

EVALUATING THE FIRST CONDITION

 TABLE 2 ( WITH S1 AS THE OUTGOING VARIABLE)
BASIS X1 X2 S1 S2 S3 bi

X2 30 2 1 1/3 0 0 6
S2 0 1 0 -1/3 1 0 2 VALUE OF
S3 0 -6 0 -5/3 0 1 0 BASIC
VARIABLE IS
Cj 28 30 0 0 0 ZERO
Zj 60 30 10 0 0
Δj = Cj - Zj -32 0 -10 0 0

NOTE THAT S3 IS A BASIC VARIABLE BUT HAS A VALUE EQUAL TO ZERO

TABLE 2 SHOWS THE OPTIMAL SOLUTION AS AS ALL Δj ARE EITHER ZERO OR

NEGATIVE
OPTIMAL SOLUTION IS: X1 = 0; X2 = 6; S2 = 2; S3 = 0
OPTIMAL VALUE OF OBJECTIVE FUNCTION: Z = 28(0) + 30(6) = 180
WITH S3 AS THE OUTGOING VARIABLE

EVALUATING THE OTHER CONDITION

 TABLE 2( WITH S3 AS THE OUTGOING VARIABLE)
BASIS X1 X2 S1 S2 S3 bi bi /X1

S1 0 18/5 0 1 0 -3/5 0 0
S2 0 11/5 0 0 1 -1/5 2 10/11
X2 30 4/5 1 0 0 1/5 6 15/2
Cj 28 30 0 0 0
Zj 24 30 0 0 6
Δj = Cj - Zj 4 0 0 0 -6

NOTE THAT S1 IS A BASIC VARIABLE BUT HAS A VALUE EQUAL TO ZERO

X1 WILL REPLACE S1
 TABLE 3( WITH S3 AS THE OUTGOING VARIABLE)
BASIS X1 X2 S1 S2 S3 bi
VALUE OF
BASIC
X1 28 1 0 5/18 0 -1/6 0 VARIABLE IS
S2 0 0 0 -11/18 1 1/6 2 ZERO
X2 30 0 1 -2/9 0 1/3 6
Cj 28 30 0 0 0
Zj 28 30 10/9 0 16/3
Δj = Cj - Zj 0 0 -10/9 0 -16/3

NOTE THAT X1 IS A BASIC VARIABLE BUT HAS A VALUE EQUAL TO ZERO

TABLE 4 IS ARRIVED AT AFTER TWO ITERATIONS. IT SHOWS THE OPTIMAL

SOLUTION AS AS ALL Δj ARE EITHER ZERO OR NEGATIVE
OPTIMAL SOLUTION IS: X1 = 0; X2 = 6; S2 = 2;
OPTIMAL VALUE OF OBJECTIVE FUNCTION: Z = 28(0) + 30(6) = 180
COMPARING THE TWO SOLUTIONS
• THE SET OF BASIC VARIABLES IS DIFFERENT

• THE VALUES OF ALL THE VARIABLES (BASIC

AND NON-BASIC) ARE THE SAME

• THE OPTIMAL VALUE OF OBJECTIVE FUNCTION

IS THE SAME