Chapter 5 .

Big -
M & Two -

phase Method .
 0 In simplex Method .
we need an initial BFS to start

computation .

0
If the LP is in feasible canonical form .
initial BFS

is provided in the initial tableau .

0 If not ,
how can we solve this LP ?

E.g Max 2- kit I

,,
amI;ᵗʰ
.
=
x2 :

St .
22 , -1 Xz -

Zz =
4 2 1

I 2 0 6
29 +222 =6

g, ,
g, ,
y, ≥ , ,

( No initial BFS ! )

construct BFS with different basis until we

get a feasible one .

let E- / %) B= cai.at >

=/
' '

)
ñ=
/ )
-

2 0 ,

at ix. =/ %;) B--

lai.at>=
/ ? ;) BFsñ=
/ ¥;)
Question Is there efficient ?
:
any more
way

form
we will solve LP not in feasible canonical
by
Big - M method and Two -

phase method .
 Big-Mmethod_
Consider a LP in standard form :

Max 2- = E' a-
(* ) s.t.AZ =3 ( 578 )

ñ≥o

Convert it to : 1 for some big M >> 0 ) .

Max Z
'
= Éñ -
¥ ,
M Kai
(** )
S.t. Asi + In Tea =3
(
auxiliary problem )
I. Tea ≥ñ .

Éoisfeasib6bC*) (É?)isfeasibleWl**#
proposition 1 .

PI 1 Exercise )

Note : For any positive Kai ,

-

Mika , '
decreases 2-
'
dramatically !
i.
opt
Sol
of ( ** ) tends to
give Kai =0 i = 1.2.3

m.PT??Yiajanop+soeJe*c--i**snasanoptsoefF#
i.
-

, .

PI ( Exercise >
 g.
E. Max 2- Kita
=

St .
22 , -1×2-23=4
(* ) =6
x, +222

Zi .
K2 , 23 ≥ 0 .

Set an
auxiliary problem :

Max 14K¢
Mass
2- =k , -122 - -

.t .
22 , -1×2-23 -1k¢ =
4
(** )
Ki +222 Tks =6

Ri .
K2 , Zz 70 . 2¢ ,
25 ≥ 0 .

( Na .
25 are
artificial variables )

-i-ioM
I. 2, Nz 23 24 25 this ratio

24 2*l 2

25 I 2 0 0 I 6 6

0
×

-
I -3M, -1-314 M O O -10M

|
-2-1 .
K, 22 23 24 25 this ratio

K, lYz-YzYz# 4

Us 0 %* Yz
-

Yz I 4 %
0,1+321-1+1=1+31--02-414
-31 rhs

43*4
Ni 22 23
.

x
,
10734T
%
°"◦_yz'#
22

◦3j|
izo#
-4-1 2 ,
22 23

se,

0 I

a-
*
( 6 85 *
6
'

opt Sol =
,
0 , with 2- =
.

Question :

If optimal Sol to (** ) has Ña* -1-8 .

What does it mean ?
 Question :

How to determine the value of M , if M -

method is

in
implemented computer ?

E.g . Min Ki

s t ER, -22 as = E. ( E is a positive constant )

(* ,
-
-
.

Xi ≥O ,
i =L .
2 .
3 .

a. =
E+a¥I =
,+ MEI

Jet 05
'
*
opt Sol : = ( 1 .
0, with 2- = 1 .

If solved using Big -

M method .

Min Ki 1- 1424

(* * ) 5. t . ER, -22 -

23 -1K¢ = E.

Xi ≥O ,
i =L .
2 .
3 . 4

T
initial BFS at =L 0,0 ,
0 , E) with 2- = ME .

In order to find the opt Sol to 1*1 ,

he must have

ME > 1 .

E > tu .

determined
M cannot be
arbitrarily .
Two-phasemethodphase.li Search for a BFS for [* )

{ phase
.

I Sol for
:
computer an
opt c* )

Max 2- = E' a-
(* ) s.t.AZ =3 ( 578 )

ñ≥o
phase 7- :

Min z
'
= 1ÉÑa = Im Kai
i =,
(* * )
St .

AZ + 7- mÑa =D
3, ña≥o .

The smallest of
'
Note :
possible value 2- = 0 .

Goal : Use Sinxex method to move the initial BFS of c** )

b- a BFS with Tea =3 .

H-sisfeasib.tl#cx-*shasanoptsol(FFa*)wit&za*--
proposition .

8 and 2- '
= 0 .

Question :

If c** ) has an
opt Sol with Ña -1-5 . What does it mean ?
By the end
of phase 1 .

if 2-1--0 initial BFS to 1*1 to phase II.

{ if
,
we
get an .
move

feasible Sol
'
2- 1=0 , (* > has no ,
stop .

In I
phase ,

LP ( * ) initial BFS obtained I

solve the original with in
phase .
E.g . Max 2- =
kit x2

St .
22 , -1 Xz -

Zz =
4

Ki +222 =6

Ri .
K2 , Nz ≥ 0 .

phase I :

Min 2-
'
=
Kat N5

a- * > S.t. 22 , -1 Xz -

23 -124 =
4

Ki +222 +25=6

Zi .
K2 , Zz ≥ 0 . 24 . 25 30 .

Question : How to solve a LP of Min

type
?

it > Convert into a Max problem .

Min 2-
'
=
Kat N5 Max -

2¢ -

25

(2) Modify optimality condition in

Simplex Method .

For optimal
0 Max problem ,
rj ≥O .
ltj .

↳,

| y÷ , r .
2-

if rj
+
µ ,}
I
,

>0
,
,
rj Kj

↑kj
=
Yoo

2- ↓

/
,
,

if rj <0 ,
Taj .
2- ↑

if 7-
rj < 0 , pick NK With rkco as entering variable .

otffffy.PT?Ytokxewn-nresoasentenhgvaiiah D- ≤ 0 .
tj optimal .
phaseI.MN 2-
'
=
Kat K5

a- * > S.t. 22 , -122 -

23 -124 =
4

Ki +222 +25=6

24 .
K2 , Nz ≥ 0 . 24 . 25 30 .

2*1--
✗ , 22 23 K¢ 25 this ratio

24 2

25 I 2 0 0 I 6 6

☐◦O
×

3 3 -

I 0 0 10

O%Yz
TI K , Nz Ks 25 this ratio

x, lYz-Y# 4

kg 0 %* Yz 1 4 93

10-%43-mgj.is/kg-
-3-1 K, 22 23 rhs

2,

0,05
'

Sol c 43 % 0 with 2-1=0

opt , , , .

( 43 ,
% ,
03T is a BFS to the original problem without

variables !
artificial
 Max 2- = K , -122

-1×2-23=4
phase St .
22 ,

Ki +222 =6

Zi .
K2 , Nz ≥ 0 .

,→Y%_
I 2 ,
22 23 rhs

Ki 10734T
m 0 I

×
0 d- (* ) Max 24 -122 .

0 0 -43 '
%

Y%%[
±
24 6

23 0 3 I

opt Sol .
E- =
( 6 ,
0,85 with 2-
*
= 6 .
Lecture 5: The Simplex Method II 5-7

Example 5.1.3.
2222 a Consider the following LP.

Min z = 2x1 + 4x2 + 7x3 + x4 + 5x5

subject to x1 + x2 + 2x3 + x4 + 2x5 = 7
x1 + 2x2 + 3x3 + x4 + x5 = 6
x1 + x2 + x3 + 2x4 + x5 = 4
x1 free, x2 , x3 , x4 , x5 0.

Since x1 is free, it can be eliminated by solving for x1 in terms of the other variables
from the first equation and substituting everywhere else. This can be done nicely using our
pivot operation on the following simplex tableau:

x1 x2 x3 x4 x5 b
1⇤ 1 2 1 2 7
1 2 3 1 1 6
1 1 1 2 1 4
2 4 7 1 5 0
Initial tableau

We select any non-zero element in the first column as our pivot element - this will
eliminate x1 from all other rows:

x1 x2 x3 x4 x5 b
1⇤ 1 2 1 2 7 (⇤)
0 1 1 0 1 1
0 0 1 1 1 3
0 2 3 1 1 14
Equivalent Problem

Saving the first row (⇤) for future reference only, we carry on only the sub-tableau with
the first row and the first column deleted. There is no obvious basic feasible solution, so we
use the two-phase method: After making b 0, we introduce artificial variables y1 0 and
y2 0 to give the artificial problem:

x2 x3 x4 x5 y1 y2 b
1 1 0 1 1 0 1 
1
0 1 1 1 0 1 3 cB =
1
0 0 0 0 1 1 0
Initial Tableau for Phase I
Lecture 5: The Simplex Method II 5-8

The cost coefficients of the artificial variables are +1 because we are dealing with a
minimization problem. Transforming (by adding the first two rows to the last row) the last
row to give a tableau in canonical form, we get

x2 x3 x4 x5 y1 y2 b
y1 1 1 0 1⇤ 1 0 1
y2 0 1 1 1 0 1 3
z 1 0 1 2 0 0 4
First tableau for Phase I

which is in canonical form. Recall that this is a minimization problem, entering variable
is chosen with positive entry (rather than negative) in the x0 -row. We carry out the pivot
operations with the indicated pivot elements:

x2 x3 x4 x5 y1 y2 b
x5 1 1 0 1 1 0 1
y2 1⇤ 2 1 0 1 1 2
z 1 2 1 0 2 0 2
Second tableau for Phase I

x2 x3 x4 x5 y1 y2 b
x5 0 1 1 1 0 1 3
x2 1 2 1 0 1 1 2
z 0 0 0 0 1 1 0
Final tableau for Phase I

At the end of Phase I, we go back to the equivalent reduced problem (i.e. discarding
the artificial variables y1 , y2 ):

x2 x3 x4 x5 b
x5 0 1 1 1 3 
1
x2 1 2 1 0 2 cB =
2
z 2 3 1 1 14
Initial Tableau for Phase II

This is transform into

Lecture 5: The Simplex Method II 5-9

x2 x3 x4 x5 b
x5 0 1 1 1 3 
1
x2 1 2⇤ 1 0 2 cB =
2
z 0 2 2 0 21
Initial Tableau for Phase II

Pivoting as shown gives

x2 x3 x4 x5 b
1 1
x5 2 0 2 1 2
1 1
x3 2 1 2 0 1
z 1 0 1 0 19
Final Tableau for Phase II

The solution x3 = 1, x5 = 2 can be inserted in the expression (⇤) for x1 giving

x1 = 7 + 2(1) + 2(2) = 1.

Thus the final solution is x⇤ = [ 1, 0, 1, 0, 2]T with z ⇤ = 19.