0 Up votes0 Down votes

0 views6 pagesindras work file

Apr 06, 2019

© © All Rights Reserved

- GATE 2013 Mathematics Previous Year Paper.pdf
- Unit 2 Matrix Algebra and Its Application
- 1
- US Federal Reserve: reliable
- Matrix Lec01 Intro
- matrix summary
- Lecnotes Econ Analiyses
- KEM BAGAN LALANG -MATHEMATICS
- Errata Leon Ed
- Summary for Linear
- Pre Mfe Nla Feb2011 Notes
- Det Eval Man
- 20120229000242SMA3013_Chapter 1
- Nonlinear Principal Component Analysis and Related Techniques
- Frequency responses for sampled-data systems
- 0207033
- Math Primer (4)
- Matrix Cookbook
- ICA Overview
- eign cha

You are on page 1of 6

first order system, what is the distance to a nearest singular

In this section we briefly explain about the concept of zeros

descriptor system that has impulsive initial conditions. The link

between impulsive initial conditions and zeros at infinity is well- at infinity and its relation with the impulsive solutions in a

known. This paper provides bounds on the minimum perturbation dynamical system. We then explain about SLRA.

required for a pair of matrices E and A such that the perturbed

A. Zeros at infinity

matrix pencil has one or more zeros at infinity. We provide closed

form solutions for the minimum value for rank one perturbations Zeros at infinity play a central role in this paper: we review

and also compare this minimum value with what is obtained by this from [?]. A polynomial matrix R(s) ∈ Rn×n [s] is said

the Structured Low Rank Approximation (SLRA) tool and other

to have a zero at a finite number λ ∈ C if rank R(λ) < n.

estimating methods.

The number of finite zeros for R(s) is given by the degree of

I. I NTRODUCTION determinant of R(s). Assume

of first order, it is known that existence of impulsive solutions

where each Ri ∈ Rn×n and Rd 6= 0. For any λ ∈ C and

for certain initial conditions are linked to the presence of zeros

R ∈ Rn×n (s) there exist square rational matrices U and V

at infinity of the corresponding matrix pencil. Just like ‘gener-

such that none of U and V have any poles or zeros at λ, and

ically’ a system is controllable, i.e. the set of uncontrollable

U RV = diag ((s − λ)µi (λ) ), the integers µi (λ) nondecreasing

systems form a thin set, it is also easy to see that generically a

in i which takes the values from 1 to n. It turns out that the

matrix pencil has no zeros at infinity. In other words, the set of

integers µi (λ) depend only on R and not on the U and V

matrix pairs (E, A) such that the pencil (sE − A) has a zero

matrices. If µ1 < 0 we say that R has (one or more) poles

at infinity is a set of measure zero in the space of all constant

at λ, the negative µi (λ)’s are called the structural pole indices

square matrices. This gives rise to the question, given a pair

at λ. If µq > 0 we say that R has (one or more) zeros at λ,

of matrices E, A, what is the minimum amount of perturbation

and the positive µi (λ)’s are called the structural zero indices at

required for the perturbed pair to have one or more zeros at

λ. The zeros/poles and their structural indices of R at infinity

infinity. This paper deals with this question and provides closed

are defined as those of Q(s) at s = 0 with Q(s) := R(λ) and

form solution for the case when the perturbation matrices ∆E

λ = 1/s.

and ∆A are each of rank one. In this paper, we use the 2-norm

The total number of poles and zeros (counted with multiplic-

of the perturbation matrices to quantify the distance of a first

ity) of P at any λ ∈ C ∪ ∞ are respectively denoted by zR (λ)

order DAE to having impulsive initial conditions.

and pR (λ) and defined by

The paper is organized as follows. The following section X X

contains preliminaries essential for this paper. Section II-E, zR (λ) := µi (λ) and pR (λ) = − µi (λ) .

µi >0 µi <0

in particular, elaborates on the link between zeros at infinity

and impulsive solutions to DAEs. Section III contains the A more direct count of the zeros and poles at infinity can

problem formulation while the main results of this paper are be obtained by counting the ‘valuations’ at ∞ for a rational

in Section IV. A few examples are investigated in Section V. matrix as elaborated in [?] and as explained briefly below. For

The well-known techniques involving Structured Low Rank a rational p(s) ∈ R(s), with p = a/b where a and b are

Approximation (SLRA) is reviewed in Section VI and this polynomials, b 6= 0, define ν(p) the valuation at ∞ of p by

section also contains a comparison of the values obtained using ν(p) := degree b − degree a and ν(0) := ∞. For a polynomial

the SLRA tool ([?]) with our closed form perturbation values matrix R ∈ Rn×m [s], define σi (R) as the minimum of the

(for rank one). Some concluding remarks are in Section VII. valuations of all i × i minors of R for each 1 6 i 6 n.

The structural indices at infinity of the rational matrix R are

*This work was supported in part by SERB, DST and BRNS, India. defined as ν1 (R) := σ1 (R), νj (R) := σj (R) − σj−1 (R) for

A. Kothyari, R. K. Kalaimani and M. N. Belur are in the Department of

Electrical Engineering, Indian Institute of Technology Bombay, India. Email: j = 2, . . . , n. This procedure gives some νi as positive integers,

{ashkothyari, rachel, belur}@ee.iitb.ac.in some negative and others zero. The absolute values of the

negative ones are summed to give the poles at infinity of R subsection. The 2-norm of a vector x ∈ Rn is the one we will

with multiplicity, while the positive ones are summed to give use throughout this paper: this is the usual Euclidean norm, and

the zeros at ∞ of R counted with multiplicity: same as the square root of the sum of all the entries’ squares.

X X Of course, for any orthogonal matrix U , the 2-norms of U x

z∞ (R) := νi and δM (R) := − νi .

νi >0 νi <0 and x are the same for each vector x ∈ Rn . Given a matrix

The matrix R is said to have no zeros at infinity if all νi 6 0. A ∈ Rn×n , there exist orthogonal matrices U and V such

Note that δM (R) is also called the McMillan degree of the that U AV = Σ, with Σ ∈ Rn×n such that Σ is a diagonal

polynomial matrix R: see [?]. matrix with its diagonal entries σ1 , . . . , σn ∈ R ordered such

We are interested in the matrix pencil sE − A which is a that σ1 > σ2 > · · · > σn > 0. These diagonal entries are called

polynomial matrix in first degree. We primarily deal with matrix the singular values and the number of nonzero singular values

pencils which are square and nonsingular as a polynomial is also the rank of the matrix A. The maximum singular value

matrix: we define them as ‘regular pencil’. A matrix pencil σ1 is also the induced 2-norm, defined as

sE − A, with E, A ∈ Rn×n is called a regular pencil if kAxk2

kAk2 := sup .

det(sE − A) 6 ≡0. Note that assuming regularity of a pencil x∈Rn \0 kxk2

implies that E and A are square. A pencil that is not regular Since there can be no ambiguity, we will use k · k2 to mean

is called singular. A necessary and sufficient condition for the either the vector 2-norm or the matrix (induced) 2-norm. For a

pencil to not have zeros at infinity is given by the following given matrix A, the Frobenius norm is defined as

proposition. v

u n X n

uX

Proposition 2.1: [?, Page 138] Suppose E, A ∈ Rn×n . The

kAkF := t a2ij .

matrix pencil sE − A has no zeros at infinity if and only if i=1 j=1

deg det(sE − A) = rank E. Pn

Note that kAk2F = i=1 σi2 and, further, for any orthogonal

B. Generalized eigenvalues matrices U1 and U2 , the Frobenius norms kU1 AU2 kF = kAkF

Closely related to (and often confused with) the notion of and also the induced 2-norms kU1 AU2 k2 = kAk2 . Rank one

to zeros at infinity is that of generalized eigenvalues at infinity matrices have the same 2-norm and the Frobenius norm, and

of the pair (E, A). For a regular pencil sE − A, a complex so does the zero matrix.

λ ∈ C such that rank (λE − A) < n is called a generalized

D. Eigenvalue sensitivity: condition number

eigenvalue of the pair (E, A). The point λ = ∞ is said to be a

generalized eigenvalue if E is a singular matrix. Multiplicities Since we deal with distance problems, and the generalized

of the generalized eigenvalues are more easily defined using eigenvalue at infinity plays a key role in this paper, we review

the so-called generalized Schur form of a pencil (E, A). See the notion of the condition number of an eigenvalue for a

[?] and [?, Page 377]. matrix. It is known that the condition number of a simple

Proposition 2.2: Suppose E, A ∈ Rn×n . Then there exist eigenvalue λ helps in an upper bound on the minimum per-

unitary matrices U and V such that both U EV and U AV are turbation required for the perturbed matrix to have a repeated

upper triangular. The generalized eigenvalues of the pair (E, A), eigenvalue at λ: see [?] and [?]. More precisely, for a matrix

with (algebraic) multiplicity, are equal to aii /eii when eii is A ∈ Rn×n , with λ ∈ C a simple eigenvalue of A, define

nonzero for i = 1, . . . , n. If eii = 0 for some i, but aii 6= 0, s(λ) := |ylT yr |, where yl and yr ∈ Cn are unit 2-norm

then these correspond to generalized eigenvalues at ∞ with that left and right eigenvectors respectively. Then there exists an

(algebraic) multiplicity. For some i, both aii and eii are zero if ∆A ∈ Rn×n such that A + ∆A has a repeated eigenvalue at λ

and only if the pencil is singular. and

s(λ)

Unlike generalized eigenvalues at infinity for the pair sE −A, k∆A k2 6 p

1 − s(λ)2

which exist if and only if rank E < n, existence of zeros at

infinity requires conditions on both E and A. Further, general- The situation for the generalized eigenvalue problem has been

ized eigenvalues at infinity is a necessary condition for zero at less investigated. We review the definition of the chordal

infinity. It is easy to see that deg det(sE − A) 6 rank E, and While the condition number of a generalized eigenvalue is

the above proposition disallows fall of the degree of determinant defined more reasonably through the ‘chordal metric’, this does

of sE − A (with respect to the rank of E) for the absence of not give a very tight upper bound on the minimum perturbation

zeros at infinity of the pencil. for a zero at infinity.

C. Singular values and the induced 2-norm E. Impulsive solutions in dynamical systems

Singular values and orthogonal matrices play a key role The zeros at infinity is more significant in systems theory as

and hence we summarize the essential background in this it is closely related to the presence of impulsive solutions in

a dynamical system. Impulsive solutions are system responses Of course, the above definition applies for generalized eigen-

which involve the dirac delta distributions and/or its derivatives. values with multiplicity too. We first note that if (E, A) has a

The following proposition makes this precise. repeated generalized eigenvalue at ∞, then (E, A) is already

Proposition 2.3: [?] Consider an autonomous singular sys- arbitrarily close to having one or more zeros at infinity, if it

tem with state space representation E ẋ = Ax, where E, A ∈ already does not have one. This can be seen as follows.

Rn×n and E singular. The free response of the system has no ??

impulsive solutions for any initial condition x(0) ∈ Rn if and IV. M AIN RESULTS

only if deg det(sE − A) = rank E.

This section contains our main results. We first state the

The following result from [?, page 1076] helps in revealing

following result that states that if rank of E is at least two

the fast subsystem more directly and is very helpful for con-

less than n, then a regular pencil sE − A either already has

ceptual purposes. See also [?, page 28].

one or more zeros at infinity or requires an arbitrarily small

Proposition 2.4: Consider E, A ∈ Rn×n and suppose r =

perturbation to have such zeros at infinity.

deg det sE − A. Then there exist nonsingular matrices M1 and

Theorem 4.1: A regular matrix pencil (E, A) with E, A ∈

M2 such that

" # " # Rn×n and rank (E) 6 n − 2 is arbitrarily close to a regular

Ir 0 As 0 matrix pencil pair (E 0 , A0 ) having a zero at infinity. More

M1 EM2 = and M1 AM2 = , (1)

0 N 0 In−r precisely, for every > 0, there exist perturbation matrices

where N is a nilpotent matrix. ∆E and ∆A such that

The significance of this result is that existence of zeros at •k∆E k22 + k∆A k22 6 , and

• s(E + ∆E ) − (A + ∆A ) has a zero at infinity.

infinity of the pair (E, A) is equivalent to the nilpotent matrix

N not being identically zero. Proof: Let the matrix E has rank n − 2. Perturbing E in the

following manner:

III. P ROBLEM F ORMULATION

σ1 0 . . . . . . . . . 0

Having noted the significance of zeros at infinity in dynami- 0 σ2 . . . . . . . . . 0

.. .. .. ..

cal systems theory, we formulate the problem that we study in ..

. . . ... . .

this paper.

.

.. . . . . . . σ

n−2 0 0

Problem 3.1: Consider E and A ∈ Rn×n . Find the mini-

.

.. . . . . . . . . .

mum value of k∆E k2 +k∆E k2 such that s(E+∆E )−(A+∆A )

1 2

has one or more zeros at infinity. 0 ... ... ... ... 0

We also deal with the situations when only E or only A Here,

are perturbed. Another important measure of the amount of 0

.

perturbation is the Frobenius norm. However, for a rank one h i ..

mtl = 0 ... 1 , mr =

matrix, the Frobenius and 2-norms are the same: rank one −2

matrices play a key role in our results. We provide explicit 1

closed form solutions for the case when ∆A and ∆E are each

hence,

of rank one. Of course, for the case when these need not be of

rank one, our values are upper bounds. h i

mtl Amr = an,1 . . . an,n−1 an,n mr

It is easy to see that

.

k∆E k22 + k∆A k22 6 k[∆E ∆A ]k22 . ⇒ ann 1 = an,n−1 2

1 an,n−1

As reviewed in Section II-E, it is often convenient to bring ⇒ =

2 an,n

the pair (E, A) to the standard form (1). However, nonsingular

Thus,

matrices M and N in general do change the 2-norm and the

1 = K × an,n−1 , 2 = K × an,n

Frobenius norms of E and A, and hence of the perturbation

matrices too. In this context, orthogonal/unitary matrices play a . So we can make as small as possible. Hence with a very small

more helpful role to bring the pair (E, A) to a convenient form value of K, we can have a zero at in finity in the perturbed

without loss of generality. We review two particular forms. The (E 0 , A) pair. If an,n−1 = 0, then 1 = 0 and 2 can be any

first one is of course the case where orthogonal matrices are arbitrary small value.

used to ensure E is diagonal, and in face, has its singular values If the matrix E has rank less than n − 2 hence with arbitrary

along the diagonal (sorted according to decreasing magnitude.) small perturbation the rank can be increased to n − 2 and the

The second one is the generalized Schur form: see [?]. same procedure can be adopted to get a zero at infinity.

If these perturbations make the perturbed pair (E 0 , A) sin- E to be r. Consider M ∈ R(n−r)×n , full row rank such that

gular, then mr is in the null space of matrix A. Hence we can M E = 0. Then the following are equivalent:

perturbe A by some small in the last column or the n − 1 1) dim(M A ker E) = n − rank E.

column such that mr is no longer in the nullspace of perturbed 2) There are no inadmissible initial conditions for this system.

A0 . Proof: Consider a singular value decomposition of any

Theorem 4.2: A singular matrix pencil (E, A) with E, A ∈ matrix, say P ∈ Rn×n , with rank r < n.

Rn×n is arbitrarily close to a regular matrix pencil having a " #

D 0

zero at infinity. In other words, for every > 0, there exist UPV = Σ = (2)

0 0n−r

∆E , ∆A ∈ Rn×n such that

where U, V ∈ Rn×n are orthogonal matrices and D is a

• s(E + ∆E ) − (A + ∆A ) is regular,

nonsingular diagonal matrix of size r. Let U be partitioned

• k∆E k22 + k∆A k22 6 , and " #

U1

• s(E + ∆E ) − (A + ∆A ) has a zero at infinity. as , where U1 ∈ Rr×n and U2 ∈ R(n−r)×n . Let V be

U2

Proof: Let the nullspaces of matrix A and matrix E be NA and h i

partitioned as V1 V2 , where V1 ∈ Rn×r and V2 ∈

NE respectively. Hence for the matrix pencil to be singular:

Rn×(n−r) . We use this decomposition, in particular that the rows

NA ∩ NE 6= {0}. of U2 are a basis of ker P T and the columns of V2 are a basis

of ker P .

First assuming the rank of E = n − 1 and intersection is (1 ⇒ 2) : We assume dim(M A ker E) = n − r and

nonempty for the right null spaces of both E and A. show that sE − A has no zeros at infinity. Without loss of

0 generality, assume E is diagonal. Since the rank of E is r. Let

. A be partitioned in accordance with nonzero and zero diagonal

NE = ..

. entries in E into: " #

1 A1 A2

A3 A4

Hence the last column of matrix A is zero. Thus by perturbing

the last column of anywhere except at the last entry i.e an,n , where A1 ∈ Rr×r and A4 ∈ R(n−r)×(n−r) . From the partitions

the perturbed pair (E, A0 ) becomes regular which has a zero at of U and V we have A4 = U2 AV2 . The rows of U2 form a basis

infinity. Similar procedure can be adopted if left null spaces are for ker E T , i.e U2 = M . Similarly columns of V2 form a basis

involved. If both left and right null space of E are contained in for ker E. Therefore dim(M A ker E) = rank A4 . It is given

the nullspace of A, then the last row and column of A are both dim(M A ker E) = n − r. This implies A4 is nonsingular.

zero. Hence perturbing simultanously the last row and column Since A4 is nonsingular and because of the structure of E it is

of A and leaving an,n = 0 the perturbed pair (E, A0 ) becomes clear that deg det (sE −A) = r. Since U and V are orthogonal

regular and having a zero at infinity. matrices the deg det (sE − A) = deg det (sE − A). Hence

Let the rank of E has rank less than n − 1 and right nullspace we have deg det (sE − A) = r = rank E. Therefore from

of E and A has a nonzero intersection. Corresponding to any proposition (4.3) there are no inadmissible initial conditions.

one column of the right nullspace of E which also corresponds (2 ⇒ 1) : We assume that there are no inadmissible initial

to a certain column in A perturb E such that rank(E) = n − 1 conditions for the singular system. Hence from proposition (4.3)

and right nullspace of E 0 contains the same column. Pertubing we have deg det (sE − A) = rank E = r. Using arguments

that particular column of A by a small amount except at the exactly like the previous part we have that if deg det (sE − A)

position corresponding to the position of one in the transpose has to be r then A4 has to be nonsingular. This implies that

of left null space of E 0 we get a perturbed pair (E 0 , A0 ) which dim(M A ker E) = n − r.

is regular and having a zero at infinity. Same procedure can be Theorem 4.4: For a regular matrix pencil (E, A) with

adopted if intersection occurs for left nullspaces of E and A or E, A ∈ Rn×n and rank (E) = n−1, the values of the minimum

if the intersection is nonzero for both left and right nullspaces norm for rank 1 perturbations on either A or E such that

E and A. In the latter case the common element on the row (E, A + ∆A ) or (E + ∆E , A) has a zero at infinity are:

and column corresponding to selected right and left nullspace Case 1:Rank 1 perturbation only on A:

of E should not be perturbed.

The following proposition characterize the condition for no σ1 0 ... 0

0 σ2 ... 0

zeros at infinity for the pencil in terms of the left and right null

.

.. .. ..

spaces of E acting on A. ..

E := . . .

Lemma 4.3: Let E ẋ = Ax, with E, A ∈ Rn×n describe an 0 . . . σn−1 0

autonomous singular state space system. Assume the rank of 0 ... ... 0

and Example

2:

a1,1 ... ... ... 9 0 0

... ... ... ... E = 0 6 0

A :=

... ... ... ...

0 0 0

...

... ... ...

2.5200 19.3340 23.9000

... ... ... an,n A = 12.5000 22.3400 6.9000

Value is an.n .

Example 3:

Case 2:Rank 1 perturbation only on E.

6 0 0

E = 0 5 0

σ1 0 ... 0

0 σ2 ... 0 0 0 0

. .. .. .. 6.9400 45.6200 −23.3000

..

E := . . .

A = 10.3200 −12.2000 4.0000

0 ... σn−1 0

12.8000 30.5000 9.2300

0 ... ... 0

VI. SLRA: F ORMULATION AND C OMPARISON

and

a1,1 ... ... a1,n In this section we will formulated the problem of SLRA.

. ..

.. ... ... . The we will relate the problem of SLRA with our problem and

. .. then compare our result with the solution obtained using the

..

A := ... ... .

. .. SLRA package provided with [LRAIM] using different initial

.

. ... ... . conditions.

an,1 ... ... an,n

A. SLRA formulation

Value of the minimum norm of the Rank 1 perturbation on the

last column of E is Structured Low Rank Approximation deals with the construc-

an,n σ1 . . . σn−1 tion of a structured low rank matrix nearest to a given matrix.

p

(an,1 σ2 σ3 . . . σn−1 )2 + . . . + (an,n−1 σ1 σ2 . . . σn−2 )2 Problem Statement: Let ω ⊂ Rn×n be a subset of matrices

having a particular structure. Let X ∈ ω, then

.

Similarly a rank 1 perturbation on the last row of E will give minimize kX − Y kw

Y

us the following value:

subject to rank Y 6 r − l, l = 1 . . . r − 1,

an,n σ1 . . . σn−1

p Y ∈ω

(an,1 σ2 σ3 . . . σn−1 )2 + . . . + (an,n−1 σ1 σ2 . . . σn−2 )2

. where k · kw denotes the weighted 2-norm. We will be using

unweighted 2-norm for solving the problem. For the structure,

V. EXAMPLES we use the following key result to find a nearest matrix pencil

Example 1: having zeros at infinity.

Consider the case when Proposition 6.1: ([?, Theorem 1]) Let P (s) ∈ Rn×n (s) and,

6 0 0 6.9400 45.6200 −23.3000 P (s) = P0 + P1 s + ... + Pd sd

E = 0 5 0 , A = 10.3200 −12.2000 4.0000

and

0 0 0 12.8000 30.5000 9.2300

Pd ... P−i+1 P−i

Then the minimum value of the norm of the Rank 1 perturbation .. ..

. .

done on the last column of E is 1.4283. If perturbation is done Ti =

Pd Pd−1

on the last row of E then the minimum value of the norm of

0 Pd

the Rank 1 perturbation is 2.3279.

Example 1: is a block Toeplitz matrix constructed from matrix coefficients

6 0 0 of A(s) and let

E = 0 5 0

ri = rank Ti − rank Ti−1 ..

0 0 0

16.9400 −1.3560 −3.5000 Polynomial matrix A(s) has no zeros at infinity if and only if

A = 20.3900 1.6900 14.0000

For our case P (s) = sE−A, and hence P0 := −A and P1 :=

E. This yields the above necessary and sufficient condition for Initial Rank 1 SLRA σn−1 (E) σ2n−1 (T0 )

p

Condi- perturbation 2

k∆A k + k∆E k2

the matrix pair (E, A) to have one or more zeros at infinity as p

" # tion 21 + 22 + ..

E A

rank − rank E 6 n − 1. Rini1 1.4283 5 5 0.2263

E 0

Rini2 1.4283 12.6741 5 0.2263

" #

E A Rini3 1.4283 42.9928 5 0.2263

Further, if rank E is n − 1, then the Toeplitz matrix

E 0 Default 1.4283 5.9389 5 0.2263

has rank at most 2n − 1. Thus if we find the nearest

" low

# rank For Example 3:

E 0 A0

matrix with the same Toeplitz structure, i.e. then σn−1 = 5, an,n = 9.2300, X = 1.4283, Y = 2.3279

E0 0

For Example 1:

E 0 − E =: ∆E and A0 − A =: ∆A .

σn−1 = 5, an,n = 9.2300 , X = 0.9452, Y = 3.2271

We next demonstrate perturbation proposed Using certain

For example 2:

examples, we will demonstrate that our perturbation is better

σn−1 = 5, an,n = 9.2300, X = 1.8363, Y = 3.1895

than answers that we get from SLRA.

VII. C ONCLUSION

B. Comparison

We related the minimum perturbation required to have ze-

In order to compare with the SLRA tool ([?]), we use the

ros at infinity for a pencil to various topics like repeated

following initial conditions in addition to the default. The values

generalized eigenvalues at infinity and Structured Low Rank

obtained with respect to these initial conditions is later below

Approximiation (SLRA). We provided explicit bounds for the

in the table.

" #

0 0 0 0 0 1.0000 case when perturbations are allowed for both E and A but the

Rini1 = perturbation matrices are each of at most rank one. We showed

0 0 0 0.0022 1.0000 0

" #

0 0 0 0 0 1.0000 using examples that the condition number defined through the

Rini2 = chordal metric gives only an upper bound which could be very

0.4472 0.4472 0.4472 0.4472 0.4472 0

conservative.

" #

1 0 0 0 0 0

Rini3 = Of course, when rank 2 or more perturbation matrices are

0 1 0 0 0 0

σ2n−1 (T0 ) σn−1 (E) Theorem 4.4 Initial SLRA allowed, then the minimum we obtained is an upper bound. A

(lower bound) (upper bound) Condition lower bound for this case is what we can get using the special

0.2227 5 0.9452 Default 8.3824

Example 1

0.2227 5 0.9452 Rini1 5

0.2227 5 0.9452 Rini2 7.5471 matrix.

0.2227 5 0.9452 Rini3 8.8892

0.7446 6 1.8363 Default 15.0833

Example 2

0.7446 6 1.8363 Rini2 17.8630

0.7446 6 1.8363 Rini3 33.6599

0.2263 5 1.4283 Default 5.9389

Example 3

0.2263 5 1.4283 Rini2 12.6741

0.2263 5 1.4283 Rini3 42.9928

p

Condi- perturbation 2

k∆A k + k∆E k 2

p

tion 21 + 22 + ..

Rini1 0.9452 5 5 0.2227.

Rini2 0.9452 19C 7.5471 0.2227

Rini3 0.9452 21C 8.8892 0.2227

Default 0.9452 1 8.3824 0.2227

p

Condi- perturbation 2

k∆A k + k∆E k 2

p

tion 21 + 22 + ..

Rini1 2.4642 6 6 0.7446.

Rini2 2.4642 17.8630 6 0.7446

Rini3 2.4642 33.6599 6 0.7446

- GATE 2013 Mathematics Previous Year Paper.pdfUploaded byprsnth
- Unit 2 Matrix Algebra and Its ApplicationUploaded byAbrha636
- 1Uploaded bySuraj Motee
- US Federal Reserve: reliableUploaded byThe Fed
- Matrix Lec01 IntroUploaded byadi.s022
- matrix summaryUploaded byApam Benjamin
- Lecnotes Econ AnaliysesUploaded byNazan Nur Kuşçuoğlu Öğünlü
- KEM BAGAN LALANG -MATHEMATICSUploaded byaikieyut
- Errata Leon EdUploaded byJohnantan Santos
- Summary for LinearUploaded byGizem Daban
- Pre Mfe Nla Feb2011 NotesUploaded bysuperauditor
- Det Eval ManUploaded byGeovane Junior
- 20120229000242SMA3013_Chapter 1Uploaded byNurul Hana Balqis
- Nonlinear Principal Component Analysis and Related TechniquesUploaded byreginaxy
- Frequency responses for sampled-data systemsUploaded byrcloudt
- 0207033Uploaded bykocayusuf13
- Math Primer (4)Uploaded byIvan Rodriguez
- Matrix CookbookUploaded byAkshay Phadke
- ICA OverviewUploaded byArunesh Mittal
- eign chaUploaded byengrfahadkhan
- CTPS-QBANK-2015Uploaded byManjunath
- MathSem3Uploaded byPooja
- Chemical Reaction Equilibrium Analysis - Theory and Algorithms (1991)Uploaded byAkash Sood
- A Note on Adjustment of Free NetworksUploaded byJajang Nurjaman
- Critical Book Report Linear AlgebraUploaded byCindy Fildza
- Resume CVTemplate_englischUploaded byBharath
- 6 EigenSingular LnvUploaded byFahrega Ridwan
- MIT18_330S12_Chapter5.pdfUploaded byupasana9
- 3.3 the Inverse of a MatrixUploaded byMrAlittle Finger
- 24.docxUploaded bypiraucatalin

- Bhagavata Pradipika#22Uploaded byBhagavata Pradipika
- 10.10 - Emotional Liquidity That is AboveUploaded byRampyare Yadav
- 13.11 - Are We Mistaking Solitary to Be LonelyUploaded byRampyare Yadav
- Pankaj Master Mridanga NotesUploaded byRampyare Yadav
- ART of TEACHING Handout Material 2 Page BGD StdUploaded byRampyare Yadav
- ProofUploaded byRampyare Yadav
- cdc-zai8Uploaded byRampyare Yadav
- cdc-zai7Uploaded byRampyare Yadav
- Hindi SB Canto 01 Part 1.pdfUploaded byRampyare Yadav
- Nagar Center Report.docxUploaded byRampyare Yadav

- The Beginnings of Alchemy ArticleUploaded byjamesilluminare
- Price List Author WiseUploaded byVipin Rai
- nm_features_20090601Uploaded byjnorine
- 211207Interlock Diagram For Unit Protection RelayUploaded byapi-3724402
- Molecular Glasses, Inc. Files Patent on New BOOster™ Technology That Increases Light Output of Solution-Coated (Inkjet) OLEDs by 150%Uploaded byPR.com
- Amazon 093 chatty report July - September 2009Uploaded byGVI_Amazon
- US Food and Drug Administration: 04n-0454-bkg0001Uploaded byFDA
- Ash HobbitUploaded byWilliam Angelico
- Www Littlebighorn Info IndianCasualtiesUploaded byUNO DOS
- 772.9 Transmission Oil and Filter ChangeUploaded byAdnan Munir
- Nsk Bearings-general CatalogsUploaded byapi-3812573
- Use of CFD Methods for Hullform ion in a Model BasinUploaded byseckin80
- 320195085 Republic Act 6969 SummaryUploaded byZarah Jeanine
- case report angelin putri gozali.docxUploaded byAngelin
- Botanical Classification of CottonUploaded bytex_hasan_014
- ALHAMBRA Expansion SummaryUploaded byRon Van 't Veer
- 13- Amines -new 2014 (1)Uploaded byilias1973
- IJCIET_08_04_239Uploaded bypadmanabhan
- Indian Sugar Industry IntroductionUploaded byjayeshvk
- Useful Leak Specification Guidelines From TQCUploaded byMirza Safeer Ahmad
- Police Log December 7, 2015Uploaded byMansfieldMAPolice
- Palafox Bares Secret Formula That Successful Cities Have _ Inquirer BusinessUploaded byReigneth Villena
- CAMS M1 IntroUploaded bydjapollyon
- pll.pdfUploaded bykurabyqld
- LCR-600 Opps ManualUploaded byJoseph Bosse
- Resume GerminUploaded bylingesh1892
- SUPERFRAC KOCH.pdfUploaded byMPilarAlfonsoGonzález
- Peningkatan Efektifitas Quality Control Project Melalui Implementasi QAQCUploaded byAnonymous z2IiUDUn
- The Chronology Protection ConjectureUploaded byKris
- Principles of StainingUploaded byWilbert Antonino Cabanban

## Much more than documents.

Discover everything Scribd has to offer, including books and audiobooks from major publishers.

Cancel anytime.