Professional Documents
Culture Documents
Textbook Admissibility and Hyperbolicity Luis Barreira Ebook All Chapter PDF
Textbook Admissibility and Hyperbolicity Luis Barreira Ebook All Chapter PDF
Barreira
Visit to download the full and correct content document:
https://textbookfull.com/product/admissibility-and-hyperbolicity-luis-barreira/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
https://textbookfull.com/product/infrared-thermography-for-
building-moisture-inspection-eva-barreira/
https://textbookfull.com/product/artificial-intelligence-and-
robotics-huimin-lu/
https://textbookfull.com/product/artificial-intelligence-and-
robotics-1st-edition-huimin-lu/
https://textbookfull.com/product/oxidation-of-c-h-bonds-lu/
Eco-Cities and Green Transport 1st Edition Huapu Lu
https://textbookfull.com/product/eco-cities-and-green-
transport-1st-edition-huapu-lu/
https://textbookfull.com/product/geospatial-technologies-for-
urban-health-yongmei-lu/
https://textbookfull.com/product/carbon-dioxide-and-
organometallics-1st-edition-xiao-bing-lu-eds/
https://textbookfull.com/product/sensing-techniques-for-food-
safety-and-quality-control-xiaonan-lu/
https://textbookfull.com/product/cognitive-internet-of-things-
frameworks-tools-and-applications-huimin-lu/
SPRINGER BRIEFS IN MATHEMATICS
Luís Barreira
Davor Dragičević
Claudia Valls
Admissibility and
Hyperbolicity
123
SpringerBriefs in Mathematics
Series Editors
Nicola Bellomo
Michele Benzi
Palle Jorgensen
Tatsien Li
Roderick Melnik
Otmar Scherzer
Benjamin Steinberg
Lothar Reichel
Yuri Tschinkel
George Yin
Ping Zhang
SpringerBriefs in Mathematics showcases expositions in all areas of mathematics
and applied mathematics. Manuscripts presenting new results or a single new result
in a classical field, new field, or an emerging topic, applications, or bridges between
new results and already published works, are encouraged. The series is intended for
mathematicians and applied mathematicians.
Admissibility and
Hyperbolicity
123
Luís Barreira Davor Dragičević
Departamento de Matemática Department of Mathematics
Instituto Superior Técnico University of Rijeka
Universidade de Lisboa Rijeka, Croatia
Lisboa, Portugal
Claudia Valls
Departamento de Matemática
Instituto Superior Técnico
Universidade de Lisboa
Lisboa, Portugal
This Springer imprint is published by the registered company Springer International Publishing AG part
of Springer Nature.
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The main objective of this book is to give a fairly broad overview of the relation
between admissibility and hyperbolicity, which goes back to the seminal work of
Perron in 1930. In particular, we present the main directions of the theory and a
few selected recent developments. We also give some applications of the relation
between admissibility and hyperbolicity, as well as pointers for further topics that
were already out of the scope of the book. Overall, the text is an invitation to the
area and we hope that it may lead to further developments.
On purpose, we avoid discussing topics that are not directly related to admissibil-
ity or that are not applications of admissibility. This allows us to highlight the main
aspects of the theory avoiding some technical complications that are not essential.
Instead, we provide detailed references to the literature for those topics that are not
discussed in the book. On the other hand, most of the material did not appear before
in book form. Moreover, much of it was rewritten on purpose for our text.
The book is dedicated to researchers as well as graduate students specializing
in differential equations and dynamical systems, with emphasis on hyperbolicity,
who wish to have a comprehensive view of the relation between admissibility and
hyperbolicity, as well as a working knowledge of its techniques. It can also be used
as a basis for graduate courses on hyperbolicity with emphasis on admissibility.
The material is divided into three parts: the core of the theory is presented
in Chapters 2–4, a few selected topics are discussed in Chapter 5, and some
applications are given in Chapter 6. In the introduction we describe in a pragmatic
manner the origins of the theory and we give a brief overview of the contents of the
book.
In Chapter 2 we present the main results of the admissibility theory in the simpler
case of exponential contractions, both for discrete and continuous time. This allows
us to give a first introduction to the relation between hyperbolicity and admissibility
without the technical complications caused by the existence of contraction and
expansion in an exponential dichotomy.
v
vi Preface
vii
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Foundations of the Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Summary of the Book. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2 Exponential Contractions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.1 Sequences of Linear Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Evolution Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3 Exponential Dichotomies: Discrete Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.1 Two-Sided Sequences of Linear Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 One-Sided Sequences of Linear Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.3 Strong Exponential Dichotomies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
4 Exponential Dichotomies: Continuous Time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.1 Two-Sided Evolution Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.2 One-Sided Evolution Families. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.3 Strong Exponential Dichotomies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
5 Admissibility: Further Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.1 Admissible Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5.2 Nonuniqueness of Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
5.3 Nonuniform Hyperbolicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6 Applications of Admissibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
6.1 Robustness of Hyperbolicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
6.2 Hyperbolic Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.3 Shadowing Property . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.4 Lyapunov Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
ix
Chapter 1
Introduction
As already noted in the preface, the main objective of this book is to give a fairly
broad overview of the relation between admissibility and hyperbolicity. In this
chapter we describe in a pragmatic manner the origins of the theory and we give
a brief overview of the contents of the book.
1.1.1 Admissibility
The study of admissibility has its origins in pioneering work of Perron in [55], where
he studied the existence of bounded solutions of the equation
n×n matrices varying continuously with t ≥ 0. It turns out that this property, usually
called the admissibility of the pair of spaces in which we take the perturbations and
look for the solutions, yields the conditional stability of the linear equation
x = A(t)x (1.2)
for the study of the robustness of hyperbolicity, which asks whether the conditional
stability persists under sufficiently small linear perturbations.
We would like to recall a particular result from [55] or, more precisely, a simple
consequence of one of the main results. Again let A(t) be n × n matrices varying
continuously with t ≥ 0.
Theorem 1.1 If Equation (1.1) has at least one bounded solution on R+ 0 for
each bounded continuous function f , then each bounded solution of the linear
equation (1.2) tends to zero when t → +∞.
As noted above, the assumption in Theorem 1.1 is called the admissibility of
the pair of spaces in which we take the perturbations and look for the solutions.
Theorem 1.1 can be regarded as the first result in the literature on the relation
between admissibility and (conditional) stability.
One can also consider the admissibility of other pairs of spaces. Although this
will be profusely illustrated along the book, and even briefly described already in
Section 1.2, we note that there is an extensive literature on the relation between
admissibility and stability, also on infinite-dimensional spaces. For a discussion of
some of the most relevant early contributions we refer to the books by Massera and
Schäffer [46] (which culminates the development started in [45]) and by Dalec’kiı̆
and Kreı̆n [25]. Related results for discrete time were first obtained by Li in [41] and
then by Coffman and Schäffer in [22]. More recently, Slyusarchuky [80] obtained
a characterization of exponential dichotomies with respect to a sequence of Banach
spaces in the particular case of the pair (∞ , ∞ ). We also refer to [40] for some
early results on infinite-dimensional spaces. For related references, we refer the
reader to [20] and for more recent work to Huy [34].
We are interested both in dynamics with discrete and continuous time. It turns out
that there are important differences between the two. This will be illustrated along
the text, but in the following paragraphs we detail already a few of these differences.
In the case of discrete time we consider a sequence (Am )m∈I of bounded linear
operators acting on a Banach space X, for some interval I ⊂ Z. Moreover, given a
sequence (ym )m∈I in X, we consider the dynamics
We say that a pair of Banach spaces (C, D) is admissible if for every (ym )m∈I ∈ C
there exists a unique sequence (xm )m∈I ∈ D satisfying (1.3).
In the case of continuous time one can consider the linear equation (1.2) for some
(bounded or unbounded) linear operators A(t) varying continuously with t in some
interval I ⊂ R. Moreover, we consider Equation (1.1) for a given perturbation f .
More generally, one can consider arbitrary evolution families of linear operators
acting on a Banach space X. In this case there are two main types of admissibility:
1.1 Foundations of the Theory 3
1. for some linear operators A(t) varying continuously with t ∈ I , we say that a pair
of Banach spaces (C, D) is strongly admissible if for each y ∈ C there exists a
unique x ∈ D absolutely continuous on each compact interval contained in I
such that (1.1) holds almost everywhere (with respect to the Lebesgue measure);
2. for an evolution family T (t, τ ) on I , we say that a pair of Banach spaces (C, D)
is weakly admissible if for each y ∈ C there exists a unique x ∈ D satisfying
t
x(t) = T (t, τ )x(τ ) + T (t, s)y(s) ds, for t ≥ τ.
τ
We emphasize that none of the two types of admissibility implies the other.
Since we deal simultaneously with discrete and continuous time, it is natural to
ask whether some results for continuous time can be deduced from corresponding
ones for discrete time or vice versa. It turns out that this is the case only for
evolution families with bounded growth and happens twice in the text: namely,
Theorems 2.5 and 4.4 can be deduced, respectively, from Theorems 2.2 and 3.4
(their counterparts for discrete time) precisely because these are the only places
where we require the bounded growth property a priori. Incidentally, both results
consider the admissibility of the pairs of spaces (Lq , Lp ) with 1 ≤ q ≤ p < ∞.
In both cases we shall give two proofs: one that is independent of the results for
discrete time and another where the result (more precisely the sufficiency of the
admissibility property for the existence of an exponential behavior) is deduced
from a corresponding statement for discrete time. This choice has to do with a few
advantages of the first approach: while the second proof is naturally shorter, the first
proof has the advantages of ensuring that the corresponding sections can be read
independently (of those for discrete time) and, more importantly, of giving a simple
prototype of the strategy in some more elaborate arguments in Chapters 4 and 5
(see the proofs of Theorems 4.6 and 5.4). Another advantage of the first proof of
Theorem 4.4 is that it gives explicitly (as much as possible) the stable and unstable
spaces of the exponential dichotomy that in the second proof are given indirectly in
terms of a corresponding dynamics for discrete time.
Another difference between discrete and continuous time has to do with the
“natural” spaces in which we take the perturbations. As a prototype, we describe
here two possible continuous time versions of the ∞ space of bounded sequences
for discrete time. Let C be the set of all bounded continuous functions f : I → R
with the supremum norm
f ∞ = sup f (t) .
t∈I
identified when they are equal almost everywhere. For some purposes this turns
out to be a more “natural” continuous time version of the space ∞ , notably to
obtain an admissibility property that is equivalent to the existence of an exponential
dichotomy (see Chapter 4).
The spaces with respect to which one can establish admissibility properties may
also differ depending on what type of admissibility we are interested on, such as
for example strong and weak admissibility in the case of continuous time, and on
what type of exponential behavior we are studying, such as for example uniform and
nonuniform hyperbolicity. These variations are considered in detail along the book.
1.1.3 Hyperbolicity
The first condition requires the invariance of the spaces E s (x) and E u (x) (the stable
and unstable spaces) under the derivative of the map, while the second condition
requires the existence of contraction and expansion along these spaces. One can
verify that E s (x) and E u (x) vary continuously with x and so when Λ is compact,
the angle between these spaces is uniformly bounded from below, that is,
inf E s (x), E u (x) > 0.
x∈Λ
This shows that condition 3 holds automatically for a compact hyperbolic set Λ.
The notion of an exponential dichotomy corresponds to consider the particular
case of a hyperbolic set
Λx = {f n (x) : n ∈ Z}
for each m ≥ n. One can easily verify that Λx is a hyperbolic set for f if and only
if there exist λ ∈ (0, 1), C > 0 and splittings
taking the projections Pn and Qn = Id−Pn associated with the splittings in (1.6).
This leads naturally to the (abstract) notion of an exponential dichotomy introduced
in Section 1.2.2.
One can consider in an analogous manner the notion of a hyperbolic set for a
flow (see Section 6.2.2) that leads similar to the (abstract) notion of an exponential
dichotomy for an evolution family introduced in Section 1.2.2. Moreover, the
former conditions can be considerably weakened, leading to more general notions
of hyperbolicity that may also occur more often (such as the notion of nonuniform
hyperbolicity; see Section 1.2.3). Namely:
1. the contraction and expansion along the stable and unstable spaces need not be
uniform, that is, the constant C in (1.4) might be replaced by a function C(x);
6 1 Introduction
2. the angles between the stable and unstable spaces may go to zero along some
trajectories when the time approaches ±∞.
In the case of discrete time (see Section 2.1), for the dynamics defined by a
sequence (Am )m∈I of bounded linear operators, with I = {n ∈ Z : n ≥ n0 } for some
n0 ∈ Z, we characterize an exponential contraction in terms of the admissibility of
the pairs of spaces
(∞ ∞ q p
0 , 0 ) and (0 , 0 ), with 1 ≤ q ≤ p < ∞.
p
Here, 0 is the set of all sequences (xm )m∈I ∈ p such that xn0 = 0.
In the case of continuous time (see Section 2.2), for the dynamics defined by an
evolution family T (t, τ ) on I , with I = [t0 , +∞) for some t0 ∈ R, we characterize
an exponential contraction in terms of the admissibility of the pairs of spaces
where Dt0 is the set of all continuous functions x ∈ C such that x(t0 ) = 0.
For the pair (Lq , Lp ) we need to assume that the evolution family has bounded
growth, while in the case of strong admissibility, for the pair (M, Dt0 ), we require
that Equation (1.1) is satisfied almost everywhere by some absolutely continuous
function x. We recall that an evolution family T (t, τ ) on I is said to have bounded
growth if there exist a, K > 0 such that
Am Pm = Pm+1 Am , whenever m, m + 1 ∈ I,
is invertible;
2. there exist λ, D > 0 such that for each m, n ∈ I with m ≥ n we have
is invertible;
2. there exist λ, D > 0 such that for each t, τ ∈ I with t ≥ τ we have
with Qt = Id − Pt and
On the other hand, in Section 3.2 we consider the case of a one-sided dynamics,
with I = N. Here there is an additional complication—the unstable space need not
be unique (in fact it is never unique when the dynamics is invertible), which causes
that the arguments need various changes.
In Section 3.3 we consider the notion of a strong exponential dichotomy for
a dynamics with discrete time on the line and we characterize it in terms of
the admissibility of the pairs in (1.7). The word “strong” means that we have
simultaneously lower and upper bounds along the stable and unstable directions.
A corresponding study for continuous time is initiated in Section 4.1, starting
with the case of a two-sided dynamics, with I = R. In particular, we characterize
an exponential dichotomy in terms of the admissibility, respectively, of the pairs of
spaces
On the other hand, in Section 4.2 we consider the case of one-sided dynamics, with
I = R+ 0 . As in Section 3.2, there is an additional complication related to the fact that
the unstable space need not be unique. Again this causes some extra complications
in the arguments and several changes are also required when passing from discrete
to continuous time.
Finally, in Section 4.3 we consider the notion of a strong exponential dichotomy
for a dynamics with continuous time and we characterize it in terms of the
admissibility of the pairs in (1.8).
with A (m, n) as in (1.5). On the other hand, in the nonuniform setting we assume
that there exist λ, D > 0 and ≥ 0 such that
The constant measures the nonuniformity of the exponential behavior. In the case
of exponential dichotomies it is necessary to consider some additional structure,
related to the existence of contraction and expansion.
It turns out that the classical notion of a uniform exponential behavior is very
stringent for the dynamics and it is of interest to look for more general types of
hyperbolic behavior. These generalizations can be much more typical (see [4]). This
is precisely what happens with the notions of a nonuniform exponential contraction
and of a nonuniform exponential dichotomy. A principal motivation for weakening
the notion of a uniform exponential behavior is that from the point of view of
ergodic theory, almost all linear variational equations on a finite-dimensional space
have a nonuniform exponential behavior. Namely, consider a flow (φt )t∈R defined
by an autonomous equation x = f (x) on Rn preserving a finite measure μ. This
means that
μ(φt A) = μ(A)
10 1 Introduction
for any measurable set A ⊂ Rn and any t ∈ R. One can show that the trajectory of
μ-almost every point x with negative (respectively nonzero) Lyapunov exponents
gives rise to a linear variational equation
1.2.4 Applications
In this chapter we present the main results of the admissibility theory in the simpler
case of exponential contractions, for both discrete and continuous time. This allows
us to give a first introduction to the relation between hyperbolicity and admissibility
without the technical complications caused by the existence of contraction and
expansion in an exponential dichotomy. The results presented here are generalized
in Chapters 3 and 4 to exponential dichotomies, respectively, for discrete and
continuous time.
One can easily verify that ∞ is a Banach space when equipped with the norm · ∞.
Moreover, the set
∞ ∞ ∞
0 = 0 (I, X) = (xm )m∈I ∈ : xn0 = 0
is a closed subspace of ∞ .
The following result gives a complete characterization of an exponential contrac-
tion in terms of the admissibility of the pair (∞ ∞
0 , 0 ) for a one-sided dynamics.
Theorem 2.1 Take I = {n ∈ Z : n ≥ n0 }. A sequence (Am )m∈I of linear operators
in B(X) has an exponential contraction on I if and only if the pair (∞ ∞
0 , 0 ) is
∞
admissible, that is, if and only if for each (ym )m∈I ∈ 0 there exists a unique
(xm )m∈I ∈ ∞0 satisfying (2.3).
Proof Assume that the sequence (Am )m∈I has an exponential contraction and take
y = (ym )m∈I ∈ ∞
0 . Let
m
xm = A (m, k)yk , for m ≥ n0 . (2.4)
k=n0
2.1 Sequences of Linear Operators 13
m
m
D
xm ≤ A (m, k)yk ≤ De−λ(m−k) yk ≤ y ∞
1 − e−λ
k=n0 k=n0
for each m ∈ I and since xn0 = 0, we conclude that x = (xm )m∈I ∈ ∞ 0 . Moreover,
it is easy to verify that property (2.3) holds. In order to establish the uniqueness of
x, we note that due to the linearity of the map y → x defined by (2.3), it suffices
to take y = 0. In this case the uniqueness of x follows directly from (2.3) since
xn0 = 0.
Now we establish the converse. Assume that for each y = (ym )m∈I ∈ ∞ 0 there
exists a unique x = (xm )m∈I ∈ ∞ 0 satisfying (2.3). We consider the linear operator
T : D(T ) → ∞ 0 defined by
x T = x ∞ + Tx ∞.
T : (D(T ), · T) → ∞
0
B(z) : (D(T ), · T) → ∞
0
for z ≥ 1 by
zvm+1 − Am vm if n0 ≤ m < n,
(B(z)v)n0 = 0 and (B(z)v)m+1 =
vm+1 /z − Am vm if m ≥ n,
(B(z) − T )v ∞ ≤ (z − 1) v T
1
D = ,
T −1 −1 − (1/t − 1)
we obtain
z ∞ ≤ z T = B(1/t)−1 y T ≤ D y ∞ = D v .
have
∗
A (m, n)v = xm = xm = t |m−n|−1 zm
D (m−n) log t
≤ t |m−n|−1 z ∞ ≤ e v
t
D (m−n0 ) log t
A (m, n0 ) ≤ A (m, n0 + 1) · An0 ≤ e An0 ,
t2
2.1 Sequences of Linear Operators 15
for m > n0 and thus inequality (2.2) also holds, taking D = max{1, D An0 /t 2 }
and λ = − log t. This completes the proof of the theorem.
q p
2.1.2 The Pairs of Spaces (0 , 0 )
Now we consider the admissibility property for other pairs of spaces. In particular,
we consider cases when the sequences (xm )m∈I and (ym )m∈I in (2.3) belong to
different spaces.
Given an interval I ⊂ Z and p ∈ [1, +∞), let p = p (I, X) be the set of all
sequences x = (xm )m∈I in X such that
1/p
x p := xm p
< +∞.
m∈I
One can easily verify that p is a Banach space when equipped with the norm · p.
Moreover, the set
p
0 = p (I, X) = (xm )m∈I ∈ p : xn0 = 0
is a closed subspace of p . We recall that if a = (an )n∈Z and b = (bn )n∈Z are
sequences of real numbers, then their convolution is the sequence a b given by
+∞
(a b)n = am bn−m , for n ∈ Z
m=−∞
1 1 1
+ = +1 (2.7)
r q p
ab p ≤ a r b q. (2.8)
q p q
only if the pair (0 , 0 ) is admissible, that is, if and only if for each (ym )m∈I ∈ 0
p
there exists a unique (xm )m∈I ∈ 0 satisfying (2.3).
Proof First observe that since p ≥ q, there exists r ≥ 1 satisfying (2.7).
Assume that the sequence (Am )m∈I has an exponential contraction and take y =
q
(ym )m∈I ∈ 0 . Moreover, let xm be as in (2.4). It follows from (2.2) that
m
m
xm ≤ A (m, k)yk ≤ D e−λ(m−k) yk = D(a b)m ,
k=n0 k=n0
where
yk if k ∈ I, e−λk if k ≥ 0,
ak = and bk =
0 if k ∈
/I 0 if k < 0.
Clearly, a ∈ q (Z, R) and b ∈ r (Z, R). Thus, by (2.7), we have a b ∈ p (Z, R),
p
which implies that x ∈ 0 . The uniqueness of x follows as in the proof of
Theorem 2.1.
q
Now assume that for each y = (ym )m∈I ∈ 0 there exists a unique x =
p q
(xm )m∈I ∈ 0 satisfying (2.3). Let T : D(T ) → 0 be the linear operator defined
p q
by (2.5) on the domain D(T ) formed by the sequences x ∈ 0 such that T x ∈ 0 .
Proceeding as in the proof of Lemma 2.1, with ∞
q
0 replaced by 0 , one can easily
verify that the operator T is closed. For x ∈ D(T ) we consider the graph norm
x T = x p + T x q.
Hence,
1/p
G · v = G · y q ≥ Gy p = A (m, n)v p
m≥n
2.1 Sequences of Linear Operators 17
and so
Lemma 2.2 There exists K ∈ N such that for all m > n0 and n ≥ K, we have
1
A (m + n, m) ≤ . (2.10)
2
Proof of the lemma Write xn = A (n, m)x for n ≥ m. To prove (2.10), take b > m
such that xb > xm /2. It follows from (2.9) that
1
xm < xn ≤ G · xm (2.11)
2 G
q
Clearly, w = (wn )n∈I ∈ 0 . Letting v = (vn )n∈I = Gw, we obtain
⎧
⎪
⎪
⎨0 n
if n < m,
vn = xn k=m 1/ xk if m ≤ n < b, (2.13)
⎪
⎩x b−1 1/ x
⎪
if n ≥ b.
n k=m k
Note that
G (b − m)1/q = G · w q ≥ v p.
b−1
1− p1
vk ≤ (b − m) v p.
k=m
1+ q1 − p1
b−1
G (b − m) ≥ vk .
k=m
18 2 Exponential Contractions
1+ q1 − p1
b−1
b−1
k
xk
G (b − m) ≥ vk =
xj
k=m k=m j =m
1
b−1
k
≥ 2
1
2 G
k=m j =m
(b − m)(b − m + 1) (b − m)2
= >
4 G 2 4 G 2
and so
(b − m)1/r < 4 G 3 ,
1
A (n, m) = A (m + kK + r, m) ≤ A (m + r, m)
2k
G
≤ ≤ 2 G e−(n−m) log 2/K .
2k
Hence, property (2.2) holds taking D = 2 G and λ = log 2/K. Finally, repeating
arguments in the proof of Theorem 2.1, one can show that property (2.2) also holds
when n = n0 .
for t, s, τ ∈ I with t ≥ s ≥ τ ;
2. given t, τ ∈ I and x ∈ X, the maps T (·, τ )x and T (t, ·)x are continuous,
respectively, on I ∩ [τ, +∞) and I ∩ (−∞, τ ].
2.2 Evolution Families 19
Unlike in the case of discrete time, one can consider different notions of
admissibility for an evolution family.
Definition 2.4 Let C and D be Banach spaces of functions x : I → X (possibly
identified when they are equal almost everywhere). We say that the pair (C, D) is
weakly admissible for an evolution family T (t, τ ) on I if for each y ∈ C there
exists a unique x ∈ D such that
t
x(t) = T (t, τ )x(τ ) + T (t, s)y(s) ds, for t, τ ∈ I with t ≥ τ. (2.15)
τ
Sometimes we need a variant of the notion of weak admissibility. The reason for
this is discussed in detail after the proof of Theorem 2.5, where one can already refer
to specific examples, but essentially it occurs because it makes no sense to fix the
value of a function at a single point (or on a set of measure zero) when the functions
are identified when they are equal almost everywhere.
Now we introduce a notion of admissibility that corresponds to consider solutions
in the classical sense (as opposed to consider weak solutions as in Definition 2.4).
Let A : I → B(X) be a continuous function on an interval I ⊂ R and let T (t, τ ) be
the evolution family associated with the linear differential equation
x = A(t)x. (2.16)
We recall that C is a Banach space when equipped with the norm · ∞ . Given
t0 ∈ I , we denote by Dt0 = Dt0 (I, X) the closed subspace of C formed by all
functions x ∈ C such that x(t0 ) = 0. Moreover, let M = M(I, X) be the set of all
locally (Bochner) integrable functions x : I → X such that
t+1
x M := sup x(s) ds < +∞,
t∈I t
identified when they are equal almost everywhere. One can easily verify that M is a
Banach space when equipped with the norm · M .
The following result gives a characterization of an exponential contraction in
terms of the notion of weak admissibility.
Theorem 2.3 Take I = [t0 , +∞). An evolution family T (t, τ ) on I has an
exponential contraction on I if and only if the pair (M, Dt0 ) is weakly admissible,
that is, if and only if for each y ∈ M there exists a unique x ∈ Dt0 satisfying (2.15).
Proof Assume that the evolution family T (t, τ ) has an exponential contraction.
Take y ∈ M and extend it to a function y : R → X by letting y(t) = 0 for t < t0 .
Let
t t
x(t) = T (t, τ )y(τ ) dτ = T (t, τ )y(τ ) dτ (2.18)
t0 −∞
D
≤ y M
1 − e−λ
Another random document with
no related content on Scribd:
Colofon
Beschikbaarheid
Dit eBoek is voor kosteloos gebruik door iedereen overal, met vrijwel
geen beperkingen van welke soort dan ook. U mag het kopiëren,
weggeven of hergebruiken onder de voorwaarden van de Project
Gutenberg Licentie in dit eBoek of on-line op www.gutenberg.org ↗️.
Metadata
H. van Brakel,
Titel:
Ing. B. O. W.
Paulus Adrianus
Info
Auteur: Daum (1850–
https://viaf.org/viaf/167261/
1898)
Aanmaakdatum 2023-09-07
bestand: 18:54:26 UTC
Nederlands
Taal: (Spelling De
Vries-Te Winkel)
Oorspronkelijke
[1875]
uitgiftedatum:
Codering
Documentgeschiedenis
2023-09-04 Begonnen.
Verbeteringen
Afkortingen
Overzicht van gebruikte afkortingen.
Afkorting Uitgeschreven
B. O. W. Burgerlijke Openbare Werken
*** END OF THE PROJECT GUTENBERG EBOOK H. VAN
BRAKEL, ING. B.O.W ***
1.D. The copyright laws of the place where you are located also
govern what you can do with this work. Copyright laws in most
countries are in a constant state of change. If you are outside
the United States, check the laws of your country in addition to
the terms of this agreement before downloading, copying,
displaying, performing, distributing or creating derivative works
based on this work or any other Project Gutenberg™ work. The
Foundation makes no representations concerning the copyright
status of any work in any country other than the United States.
1.E.6. You may convert to and distribute this work in any binary,
compressed, marked up, nonproprietary or proprietary form,
including any word processing or hypertext form. However, if
you provide access to or distribute copies of a Project
Gutenberg™ work in a format other than “Plain Vanilla ASCII” or
other format used in the official version posted on the official
Project Gutenberg™ website (www.gutenberg.org), you must, at
no additional cost, fee or expense to the user, provide a copy, a
means of exporting a copy, or a means of obtaining a copy upon
request, of the work in its original “Plain Vanilla ASCII” or other
form. Any alternate format must include the full Project
Gutenberg™ License as specified in paragraph 1.E.1.
• You pay a royalty fee of 20% of the gross profits you derive from
the use of Project Gutenberg™ works calculated using the
method you already use to calculate your applicable taxes. The
fee is owed to the owner of the Project Gutenberg™ trademark,
but he has agreed to donate royalties under this paragraph to
the Project Gutenberg Literary Archive Foundation. Royalty
payments must be paid within 60 days following each date on
which you prepare (or are legally required to prepare) your
periodic tax returns. Royalty payments should be clearly marked
as such and sent to the Project Gutenberg Literary Archive
Foundation at the address specified in Section 4, “Information
about donations to the Project Gutenberg Literary Archive
Foundation.”
• You comply with all other terms of this agreement for free
distribution of Project Gutenberg™ works.
1.F.
Most people start at our website which has the main PG search
facility: www.gutenberg.org.