• Tidak ada hasil yang ditemukan

Module 2 Topic 2:

N/A
N/A
Protected

Academic year: 2024

Membagikan "Module 2 Topic 2:"

Copied!
4
0
0

Teks penuh

(1)

Module 2

1

Topic 2:

Relationship between extreme points and corresponding BFS:

We now relate the extreme point of the feasible set with BFS of the system of constraint describing the feasible set.

Theorem 2.1: Every extreme point of a set of feasible solution of the LPP (P) max z = cTx

subject to Ax = b, x≥0 is a BFS and vice-versa.

Proof: Let x be a BFS of LPP, i.e, x = [xB,0] ,where 0 is a null vector of n−m component and xB is the vector of m basic variable in BFS then xB = B−1b. Now to the contrary let x be not an extreme point of the feasible set S of (P). Then there exist x1, x2 ∈S, x1 6=x2 and λ∈(0,1) such that

x = (1−λ)x1+λx2

"

xB 0

#

= (1−λ)

"

u1 v1

# + λ

"

u2 v2

# ,

where u1 and u2 respectively are the vector values in x1 and x2 corresponding to the basic variablex and v1, v2 are vectors of those values variables which are zero in x. We have,

λB = (1−λ)u1+λu2 0 = (1−λ)v1+λv2

The second equality yieldv1 = 0, v2= 0 , sincev1, v2 ≥0 and λ∈(0,1). Thus x1 =

"

u1

0

#

, x2 =

"

u2

0

#

Nowx1, x2 ∈S, so,

Bu1=b=Bu2

implying

B−1(Bu1) = B−1(Bu2) i.e u1 = u2.

But this contradicts the fact that x1 6=x2. Sox is an extreme point of S. We next prove that ifx∈S is an extreme point of the feasible set S of (P) thenx is a BFS.

Suppose firstkvariable in x are non-zero,k≤m. That is, x = (x1, x2, . . . xk,0, . . . ,0)T ∈S , and

k

X

i=1

aixi= b.

c

Copyright Reserved IIT Delhi

(2)

2

We will prove that the set {a1, . . . , ak} in <m is a linearly independent set. Let us suppose that this set is linearly dependent. Then there exist scalars λ1, . . . , λk such that

k

X

i=1

λiai= 0 (1)

with at least one λi 6= 0.

Define µ = min n

xi

i|, λi6= 0o

.Then µ >0 . Next chooseε, 0< ε < µ. Then xi+ελi>0 ∀i= 1,2,3, . . . , k

xi−ελi>0 ∀i= 1,2,3, . . . , k

To explain it , if λi > 0 then obviously xi+ελi > 0 and as ε < µ ≤ xλi

i. So, xi −ελi > 0 . Similarly if λi <0, thenxi−ελi >0 and as ε < µ≤ −λxi

i , soxi+ελi >0.

Define a vector λ whose first k-component are λ1, . . . , λk , and rest n−m components are zero.Take

ˆ

x = x+ελ∈ <n

˜

x = x−ελ∈ <n.

Then ˆx ≥ 0. Also as Aλ = 0 in view of (1), so, Aˆx = A˜x = Ax = b, giving ˆx,x˜ ∈ S.

Moreover,ε >0 and at least oneλi6= 0 so ˆx6= ˜x , withx= 12(ˆx+ ˜x). But this contradicts that x is an extreme point of S. Hence the set {a1, a2, ....ak} is linearly independent set. Ifk =m, then by taking a bases matrix as

B = [a1, a2, . . . ak]

We have x is a BFS of LPP. If k < m, then since rank(A) = m so the set {a1, a2, . . . ak} can be extended by adding k−m columns from A to form basis of<m. Hence x is a BFS of LPP.

Only thing to realize is that in this case xis a degenerate BFS.

Example 2.5: Consider the following LPP.

max x1+x2

subject to x1−x2 ≥0 x1+x2 ≤4 x1, x2≥0

c

Copyright Reserved IIT Delhi

(3)

3 The feasible set and the three extreme points are shown in the Fig 1.1.

C( 22, )

B(4,0) >

<

A(0,0)

x1

x2

Fig 1.1

For A (0,0) the BFS is (0,0,0,4) while for B (4,0) and C (2,2), the BFS are (4,0,4,0) and (2,2,0,0) respectively. Note that while talking of BFS, the constraints of LPP are converted to equations by introducing additional variables as

x1−x2−x3 = 0

x1+x2+x4 = 4 (2)

x1, x2, x3, x4 ≥ 0

This example also shows that for an extreme point A(0,0) there are three bases namely B1 = {x3, x4}, B2 = {x1, x4} and B3 = {x2, x4}. This is because the BFS (0,0,0,4) is a degenerate BFS. Also, if we take variablesx1= 0, x4 = 0 then the system becomes

−x2−x3 = 0 x2 = 4

which does not yield a BFS asx3 =−4<0.So, among the five possible BFS of system (2), the other two yield extreme points B and C.

Remark 2.1: If the LPP has a degenerate BFS then more than one basis correspond to the same extreme point of the feasible set LPP. This fact is important and should be kept in mind as later on we would be using it.

c

Copyright Reserved IIT Delhi

(4)

4

Example 2.6: Consider the following LPP

min z=x1+x2 subject tox1+ 2x2 ≤0 x1+x2 ≥0 x2 ≤0 x1, x2 ≥0 The feasible set is shown in the figure.

D: 2, )( 4

B:(1,0)

x

1

x

2

E:(0,4)

0 C:(10,0)

A:(0,1)

Fig 1.2

It has five extreme points A, B, C, D, E. Now to see their correspondence with BFS, we convert the system of inequalities representing the constraints to equation by introducing additional variables in them. Consider

x1+ 2x2+x3 = 10 x1+x2−x4 = 1

x2+x5 = 4 x1, x2, x3, x4, x5 ≥ 0 For A(0,1), the BFS is (0,1,8,0,3)

For B(1,0), the BFS is (1,0,9,0,4) For C(10,0), the BFS is (10,0,0,9,4) For D(2,4), the BFS is (2,4,0,5,0) For E(0,4), the BFS is (0,4,2,3,0).

c

Copyright Reserved IIT Delhi

Referensi

Dokumen terkait

The application is able to manage topic-map document according to standard format of XML Topic Map (XTM) including topic, occurrence, and association.. Topic- map

The objective of this research is to measure under achiever motivation level towards Teaching and Learning T&L Trigonometry II topic using GeoGebra Learning Module GLM.. This research

Item Percentage Content validity coefficient Expert consent 1 The content of this module meets its target population 90% .90 Accepted 2 The content of this module is

An allocation is feasible if the shared good production constraint 1 is satisfied for all m ∈ M, the constraint on the production of private goods 2 is satisfied, the sharing technology

The table shows that the topic model are classified based on two scenarios supervised and unsupervised and the methodology used for topic models were bag of words and sequence of words

CONCLUSION AND SUGGESTION Based on the results of research that has been carried out on the implementation of e-learning module based of STEM at the Unsyiah Physics Education Study

In this study using the LDA method as an algorithm to produce topic modeling, each topic similarity, and visualization of topic clusters from the tweet data generated as many as 4

Conclusion This research succeeded in developing one of the innovative learning media products in the form of STEM-based science e-modules on the human excretory systems topic for