You are on page 1of 19

Chapter 6 Special cases in

simplex Method .
Any LP be ① infeasible

§
0 can

unbounded feasible region with (without


opt
Ye
-

③ multi
opt sols
p

Question :

can we detect these properties from the


simplex tableau ?

o we will disuss the


following 4
special cases :

① No feasible Sol .

f¥ opt Sol
② unbounded feasible I
region
opt Sol .

③ multiple opt sols


④ degeneracy .
① LP has no feasible Sol .

method
This LP will be solved by either Big -
M

or Two -

phase method .

EG . MAX 2- = 2K , -122

St .
- N, -12132

2. it Ns E l

Ki 30 .

Convert it into standard forms .

MAX 2- = 2K , -122

St .
- N, -121 -
Kz = 2

2e Ns
, t +24=1

Hi 30 .
I -_ 1.2 . 3.4 .

H) Solve it
using Big -

M method .

MAX 2- = 2K , -122 -
M 25 .

St .
- N, -12k -
Zz +25=2
(* * )
Kitna +24 = I Ns :
artificial variable .

Ni 30 .
I -_ 1.2 .
3. 4.5

,,*o
TI .
Ki Na KS 204 Us the patio
-

Ns
-
l l -
I 0 I 2 2/1=2

µ y, =
,

-2 -

I O O M O

-2in -
I -
M M O O -2M
,,o#
-2L Ki Ks Hs Uk Ks rhs
-

25 -2 O -
l -

l l l

z,

ITZM O M HM O l M
-
-

opt Sol with 25 to

original problem is infeasible .

② Solve it
using Two-phase method
Nih
phase 25

St .
- N, -12<-23 +25=2
(* * )
2e Ns
, t +24 = I

Ni 30 .
I -_ 1.2 . 3.4 .

rhs ratio

,,*
I ki Nz Ks 24 Us

25 -

l l -
I 0 I 2 2

µ, ,

O O O O
-

I O
G -

l l
-
I O O 2

,,c÷
It KiN2K3k4N5t#
Ks -2 O -
I -

l l l

a,

-2 O -

I -
I 0 I

I ends
phase with 25 to .

original problem is infeasible .


② unbounded feasible region .

-0
If be increasedone variable can to an
arbitrarily large number

without
violating any constraint ,
the feasible region is unbounded

in at least that direction .

0 Furthermore ,
for a Max problem ,

if be T
upper bound opt Sol
{
Z can without an No .

from
if Z is bounded above I
opt
Sol .

E.g . Max Z =
22Gt Nz

Sit . N, -

Nz E IO
Z

2x ,
-
ka E 40

N, .
Na 30 .

Convert it into standard form .

Max Z = 2K , t Nz

Sit . N, -

Nz +23 = 10

2N K2
1-24=40
-

R, .
Na 30 .
Nz 24 30
, ,

z,,o,|#
It K,NzKsK¢r# ratio

Kz I * -
l l O lo lo

ya zo
,

-
2 -
I O O O
It 21222324M€
2e ,

o,*→ I l l O 10
-

µ,

O -3 2 O 20

o.fi#/m--m7.IIso
TI x,KzN3K¢t#
/ O I I 30

I.
K, Nz
-
= -

24+30

O O -4 3 80 2- =
4223

Not an
opt
Sol ,
but we cannot
perform any further pivot operation .

K, Nz 24+30

{
= -

22=223-24+20

z 4223-324*80
-

T
F S> 0 ,
let Nz = S . 24=0 . then C 8+30 ,
28 -120,8 .
O )

is a feasible Sol .
unbounded feasible region .

Also ,
Z = IS Tso .
No finite opt Sol .

For a Max
problem ,
if there is some rkco .
and Yik EO Citi "
- m)

for some nonbasic variable Kk ,


then this LP has no finite opt
Sol

¥÷¥÷
.

.
E.G . Max 2- =
6K , -222

S.T. 2K , -

R2 EZ

K, £4

Hi ,
Nz 30

Convert it into standard form ,

Max 2- =
6K , -222

S.T. 2K , -

Nz Tks =2

*I 1-24=4

Hi ,
Nz 23 , Ny 30

,c
I KiN2N3k¢r#
Nz 2¥ -

l l O 2

-
G 2 O O 0

0Yz*_yz#
It KiK2N3K¢r#
K, I 72 42 0 I

24

O -
I 3 o 6

017€
It kikzksk.ch#
I O O 1
2e , 4

22

O O 2 2 12
opt Sol .
I# ( 4 .
6 ,
O ,
05 with 2-
*
= 12 .

Question .

If we don't plot its feasible region ,


we
may not realize

that unbounded Can identify the unbounded ness


it 's . we
of
feasible region from the tableau ?
t÷+u¥baaii%%%ibasicuariauesgu*nye
If
then the

moreover ,
feasible

rj =
region

Zj g-
-
co
is

.
unbounded

then ,
this
.

LP has no finite opt Sol .

PI .
let IB be the B. v .

of the current BFS with basis matrix B .

B =

( tf bfz ,
.
. . .

bfm ) bi is the ith Col of B .

KB ; B. v
is the ith .

accordingly .

BBB = I ¥ ,
NB -

.
Tsi =P .

let AJ be the jth Col in A associated . with


Nj .
then .

B-' AJ =
5; Ij =
BE =
Ey Ii

For all 0>0 ,

b- =
EI RBI Toi
,
-

oaf +
Oaf
=
Ei 2B ,
.
-
Toi -

O C
II. yijbi ) to
aj
=

¥ ,
( NB -

.
-

Oyij ) b- i +09J

we obtain a new feasible Sol of Mti nonzero variables .

Where Rbi =
NB
Oyij

{
-
-

nj = O

Nia = 0 for other K .

Since the value


of Nj which is O can be
, ,
set
arbitrarily large ,

indicating that the feasible region is unbounded in the


Nj direction .
If moreover ,
rj co ,

Z
'
=
Em CBI KB,
'

;
t
Cj Nj
'

=
ET CB,
-

,
( Kisi -

Oyij ) t
Cj O

=
IT ,
Cos -243
,
.

.
-

0¥ ,
CB -

.
bij t
G-
-

£5343 0£55 g- O
= -

+
-

Z O
Zjt Cj O
-

O C
z
Cj )
=
Zj
-
-

Z
Orj
= -

Since '

tjao ,
as O T to th , Z T to tis .

# .

Question what Min ?


happens in a
problem
Feasible region is unbounded :

Moreover, no finite Sol :


opt
#
③ infinite many opt
sols

E.g
. Max 2- =
42 , -11422

Sit .
2K , -1722 £21

72 , + 222 £21

Hi ,
Nz 30

Convert it into standard form .

Max 2- =
42 , -11422

Sit .
2K , -1722 + Nz = 21

72 , -1222 +24=21
Hi ,
H2 23 , Hq 30

K,N2N32e4r#

yzo#
I ratio

as 2 7 * I 0 21 3

µ, %
-4 -

14 O O O

4470-471/1500
II kikzksk.co#
K2 47 I 47 O 3

see,

O 2 O 42

Sol II ( O 3 O 155' with Z*= 42


opt
-_
, , , .
A further attempt to do pivot
operation at 454 .

It this

10-445745/730
.
K, Nz Kz 24

01745-445%2
Kz

O 2 O 42

a
different set
of opt Sol IF = ( 73 ,
73 ,
o, o ,
T

with It = 42 .

All sols dat t Ctd ) IF V-xe-lo.is


opt
.

:
.

If in the
opt tableau . I a nonbasic variable
Nj with

rj =D ,
then there are
infinite number
of opt sols .
④ Degeneracy .

O .

degenerate BFS :
a BFS with one or more B.v .
at o levee .

o redundant
information causes
degeneracy .

B. v
if rank CA ) =m ,
in .
n - m N B. v
. .

Each boundary of the feasible region corresponds to one N B. V


. .

n -
m boundaries are
sufficient to determine a BFS

But if there are more than n -


m boundaries passing through
the same extreme
pts .

multiple sets of n -
m boundaries correspond to the same

extreme
pt .

F.G .
Max 2- = 2K , t KL ← Meo

St .
4N , -1322 E 12

42, t 22 ES ←
Us =D

4K , -

Na Ef

-0 ← 23=0
q,
K , . Nz 30 .

convert it into standard form . as -


- o

3 B. v

I
.

Max Z = 2K , t KL rank CA ) =3 2 N .
B. v .

St .
4N , -1322 tug = 12 2 lines are sufficient to determine
42, t 22 +24 = 8 1 extreme pt .

4K , -

Na tag =8 But at ( 2.0 ) there are 3 lines .

Ki . 22 , Rs , Na .
B- 30 .
3 different sets of B. v .

information redundancy , .
Example.

Max x0 = 2x1 + x2
Subject to 4x1 + 3x2  12
4x1 + x2  8
4x1 x2  8
x1 , x2 0

x1 x2 x3 x4 x5 b
x3 4 3 1 0 0 12
x4 4⇤⇤ 1 0 1 0 8
x5 4⇤ 1 0 0 1 8
x0 2 1 0 0 0 0

(⇤) . & (⇤⇤)

x1 x2 x3 x4 x5 b x1 x2 x3 x4 x5 b
x3 0 4 1 0 1 4 x3 0 2⇤⇤ 1 1 0 4
x4 0 2⇤ 0 1 1 0 x1 1 1/4 0 1/4 0 2
x1 1 1/4 0 0 1/4 2 x5 0 2 0 1 1 0
x0 0 3/2 0 0 1/2 4 x0 0 1/2 0 1/2 0 4

Degenerate Vertex {x4 = 0 and basic} Degenerate Vertex {x5 = 0 and basic}
# #
x1 x2 x3 x4 x5 b x1 x2 x3 x4 x5 b
x3 0 0 1 2 1⇤ 4 x2 0 1 1/2 1/2 0 2
x2 0 1 0 1/2 1/2 0 x1 1 0 1/8 3/8 0 3/2
x1 1 0 0 1/8 1/8 2 x5 0 0 1 2 1 4
x0 0 0 0 3/4 1/4 4 x0 0 0 1/4 1/4 0 5

Degenerate Vertex {x2 = 0 and basic}


# /

x1 x2 x3 x4 x5 b
x5 0 0 1 2 1 4
x2 0 1 1/2 1/2 0 2
x1 1 0 1/8 3/8 0 3/2
x0 0 0 1/4 1/4 0 5
aopertiaonratiama.EIY.ae?degenerateBFsatterpiot

xBj§_..---bs----|p#
Ks ratio
#
:

KB; bis
pi BYy.es
- r - - - - - -
. . -
- - - -
- -

Bibs .

Vs Yoo

pivot at y is i

:÷÷÷ti÷÷:l:÷:.÷ t
B. V
KBJ is a
degenerate
.

O tableau to the
3 all
correspond same extreme
pt ,
so , there is no

change in
obj value .

Question : Is it
possible that the simplex method cycles through
these 3 or part of these tableau s forever ?
rule to select :

variable
entering :
Most negative
{
.

in Nuh ratio
leaving variable :
then there is a tie ,

pick the one with the smallest index


Example . .

Max x0 = 20x1 + 12 x2 6x3 + 34 x4 L

Subject to x1  2
8x1 x2 + 9x3 + 14 x4  11
12x1 12 x2 + 3x3 + 12 x4  24
x2  1
x1 , x2 , x3 , x4 0

Tableau 0: Tableau 1:
x1 x2 x3 x4 x5 x6 x7 x8 b x1 x2 x3 x4 x5 x6 x7 x8 b
x5 1⇤ 0 0 0 1 0 0 0 2 x1 1 0 0 0 1 0 0 0 2
x6 8 1 9 1/4 0 1 0 0 16 x6 0 1 9 1/4⇤ 8 1 0 0 0
x7 12 1/2 3 1/2 0 0 1 0 24 x7 0 1/2 3 1/2 12 0 1 0 0
x8 0 1 0 0 0 0 0 1 1 x8 0 1 0 0 0 0 0 1 1
x0 20 1/2 6 3/4 0 0 0 0 0 x0 0 1/2 6 3/4 20 0 0 0 40

Tableau 2: Tableau 3:
x1 x2 x3 x4 x5 x6 x7 x8 b x1 x2 x3 x4 x5 x6 x7 x8 b
x1 1 0 0 0 1 0 0 0 2 x1 1 3/8 15/4 0 0 1/2 1/4 0 2
x4 0 4 36 1 32 4 0 0 0 x4 0 8⇤ 84 1 0 12 8 0 0
x7 0 3/2 15 0 4⇤ 2 1 0 0 x5 0 3/8 15/4 0 1 1/2 1/4 0 0
x8 0 1 0 0 0 0 0 1 1 x8 0 1 0 0 0 0 0 1 1
x0 0 7/2 33 0 4 3 0 0 40 x0 0 2 18 0 0 1 1 0 40

Tableau 4: Tableau 5:
x1 x2 x3 x4 x5 x6 x7 x8 b x1 x2 x3 x4 x5 x6 x7 x8 b
x1 1 0 3/16 3/64 0 1/16 1/8 0 2 x1 1 0 0 0 1 0 0 0 2
x2 0 1 21/2 1/8 0 3/2 1 0 0 x2 0 1 0 5/2 56 2⇤ 6 0 0
x5 0 0 3/16⇤ 3/64 1 1/16 1/8 0 0 x3 0 0 1 1/4 16/3 1/3 2/3 0 0
x8 0 0 21/2 1/8 0 3/2 1 1 1 x8 0 0 0 5/2 56 2 6 1 1
x0 0 0 3 1/4 0 2 3 0 40 x0 0 0 0 1/2 16 1 1 0 40

Tableau 6: Tableau 7:
x1 x2 x3 x4 x5 x6 x7 x8 b x1 x2 x3 x4 x5 x6 x7 x8 b
x1 1 0 0 0 1 0 0 0 2 x1 1 0 0 0 1 0 0 0 2
x6 0 1/2 0 5/4 28 1 3 0 0 x6 0 1 9 1/4⇤ 8 1 0 0 0
x3 0 1/6 1 1/6 4 0 1/3⇤ 0 0 x7 0 1/2 3 1/2 12 0 1 0 0
x8 0 1 0 0 0 0 0 1 1 x8 0 1 0 0 0 0 0 1 1
x0 0 1/2 0 7/4 44 0 2 0 40 x0 0 1/2 6 3/4 20 0 0 0 40
To avoid
-

cycling , bring th 24 and remove Ky .

EA

)
Tf .

NiNzN3K4K5N6N7K8r#
2e , I O O O l O O O 2

3/4 5/2
'
O O -2 I Yz O O
-

ay
-

µ, o , g , *, o z o o

Ng O Itt O O O O O l l

0-5/42420203/2047

|
TI NiN2N3K4K5K6K7K8r#
2e , I O O O l O O O 2

% 3/4 44
'
O O O -2 I
-

Yz
26

µ, o o g , se, o z , ,

Zz O l O O O O O l l

00420203kt ,

opt Sol !
Bland 's rule
=
To
"
avoid cycling ,
use .
to select

entering and departing variables .

① entering variable :
among all N B. v-
.

Nj with
rjco .

Select the one with the smallest index .

② departing variable : when there is a tie in Min ratio test,

select the one with the smallest index .

By Bland 's rule is It entering variable as

{
:
, .
.

variable 28
departing :
Example .

Max x0 = 20x1 + 12 x2 6x3 + 34 x4


Subject to x1  2
8x1 x2 + 9x3 + 14 x4  11
12x1 12 x2 + 3x3 + 12 x4  24
x2  1
x1 , x2 , x3 , x4 0

Apply the Bland’s rule:


Tableau 0:
x1 x2 x3 x4 x5 x6 x7 x8 b
x5 1⇤ 0 0 0 1 0 0 0 2
x6 8 1 9 1/4 0 1 0 0 16
x7 12 1/2 3 1/2 0 0 1 0 24
x8 0 1 0 0 0 0 0 1 1
x0 20 1/2 6 3/4 0 0 0 0 0

Tableau 1:
x1 x2 x3 x4 x5 x6 x7 x8 b
x1 1 0 0 0 1 0 0 0 2
x6 0 1 9 1/4 8 1 0 0 0
x7 0 1/2 3 1/2 12 0 1 0 0
x8 0 1⇤ 0 0 0 0 0 1 1
x0 0 1/2 6 3/4 20 0 0 0 40

Tableau 2:
x1 x2 x3 x4 x5 x6 x7 x8 b
x1 1 0 0 0 1 0 0 0 2
x6 0 0 9 1/4 8 1 0 1 1
x7 0 0 3 1/2⇤ 12 0 1 1/2 1/2
x2 0 1 0 0 0 0 0 1 1
x0 0 0 6 3/4 20 0 0 1/2 81/2

Tableau 3:
x1 x2 x3 x4 x5 x6 x7 x8 b
x1 1 3/8 15/4 0 0 1/2 1/4 0 2
x6 0 0 15/2 0 2 1 1/2 3/4 3/4
x4 0 0 6 1 24 0 2 1 1
x2 0 1 0 0 0 0 0 1 1
x0 0 0 21/2 0 2 0 3/2 5/4 165/4
optimal solution ⌅

You might also like