Hamiltonian Perturbation Theory (And Transition To Chaos) 4515

Hamiltonian Perturbation Theory (and Transition to Chaos)
Hamiltonian Perturbation Theory

(and Transition to Chaos)
HENK W. BROER1 , HEINZ HANSSMANN2
1 Instituut voor Wiskunde en Informatica,
Rijksuniversiteit Groningen, Groningen,
The Netherlands
2 Mathematisch Instituut, Universiteit Utrecht,
Utrecht, The Netherlands
Article Outline
Glossary
Denition of the Subject
Introduction
One Degree of Freedom
Perturbations of Periodic Orbits
Invariant Curves of Planar Dieomorphisms
KAM Theory: An Overview
Splitting of Separatrices
Transition to Chaos and Turbulence
Future Directions
Bibliography
dle third from a closed interval) designates topological

spaces that locally have the structure Rn Cantor dust
for some n 2 N. Cantor families are parametrized by
such Cantor sets. On the real line R one can dene
Cantor dust of positive measure by excluding around
each rational number p/q an interval of size
2
;
q
> 0; >2:
Similar Diophantine conditions dene Cantor sets

in Rn . Since these Cantor sets have positive measure
their Hausdor dimension is n. Where the unperturbed system is stratied according to the co-dimension of occurring (bifurcating) tori, this leads to a Cantor stratication.
Chaos An evolution of a dynamical system is chaotic if its
future is badly predictable from its past. Examples of
non-chaotic evolutions are periodic or multi-periodic.
A system is called chaotic when many of its evolutions
are. One criterion for chaoticity is the fact that one of
the Lyapunov exponents is positive.
Diophantine condition, Diophantine frequency vector
A frequency vector ! 2 Rn is called Diophantine if
there are constants > 0 and > n 1 with
Glossary
Bifurcation In parametrized dynamical systems a bifurcation occurs when a qualitative change is invoked by
a change of parameters. In models such a qualitative
change corresponds to transition between dynamical
regimes. In the generic theory a nite list of cases is obtained, containing elements like saddle-node, period
doubling, Hopf bifurcation and many others.
Cantor set, Cantor dust, Cantor family, Cantor stratication Cantor dust is a separable locally compact space
that is perfect, i. e. every point is in the closure of its
complement, and totally disconnected. This determines Cantor dust up to homeomorphisms. The term
Cantor set (originally reserved for the specic form of
Cantor dust obtained by repeatedly deleting the mid-
jhk; !ij

jkj
for all k 2 Z n nf0g :
The Diophantine frequency vectors satisfying this condition for xed and form a Cantor set of half lines.
As the Diophantine parameter tends to zero (while
remains xed), these half lines extend to the origin.
The complement in any compact set of frequency vectors satisfying a Diophantine condition with xed
has a measure of order O( ) as # 0.
Integrable system A Hamiltonian system with n degrees
of freedom is (Liouville)-integrable if it has n functionally independent commuting integrals of motion.
Locally this implies the existence of a torus action,
a feature that can be generalized to dissipative sys-
4515
4516
tems. In particular a mapping is integrable if it can be

interpolated to become the stroboscopic mapping of
a ow.
KAM theory KolmogorovArnoldMoser theory is the
perturbation theory of (Diophantine) quasi-periodic
tori for nearly integrable Hamiltonian systems. In the
format of quasi-periodic stability, the unperturbed and
perturbed system, restricted to a Diophantine Cantor
set, are smoothly conjugated in the sense of Whitney.
This theory extends to the world of reversible, volumepreserving or general dissipative systems. In the latter
KAM theory gives rise to families of quasi-periodic attractors. KAM theory also applies to torus bundles, in
which case a global Whitney smooth conjugation can
be proven to exist, that keeps track of the geometry. In
an appropriate sense invariants like monodromy and
Chern classes thus also can be dened in the nearly
integrable case. Also compare with Kolmogorov
ArnoldMoser (KAM) Theory.
Nearly integrable system In the setting of perturbation
theory, a nearly integrable system is a perturbation of
an integrable one. The latter then is an integrable approximation of the former. See an above item.
Normal form truncation Consider a dynamical system
in the neighborhood of an equilibrium point, a xed or
periodic point, or a quasi-periodic torus, reducible to
Floquet form. Then Taylor expansions (and their analogues) can be changed gradually into normal forms,
that usually reect the dynamics better. Often these
display a (formal) torus symmetry, such that the normal form truncation becomes an integrable approximation, thus yielding a perturbation theory setting. See
above items. Also compare with Normal Forms in
Perturbation Theory.
Persistent property In the setting of perturbation theory,
a property is persistent whenever it is inherited from
the unperturbed to the perturbed system. Often the
perturbation is taken in an appropriate topology on the
space of systems, like the Whitney Ck -topology [72].
Perturbation problem In perturbation theory the unperturbed systems usually are transparent regarding their
dynamics. Examples are integrable systems or normal
form truncations. In a perturbation problem things are
arranged in such a way that the original system is wellapproximated by such an unperturbed one. This arrangement usually involves both changes of variables
and scalings.
Resonance If the frequencies of an invariant torus with
multi- or conditionally periodic ow are rationally
dependent, this torus divides into invariant sub-tori.
Such resonances hh; !i D 0, h 2 Z k , dene hyper-
planes in !-space and, by means of the frequency

mapping, also in phase space. The smallest number
jhj D jh1 j C C jh k j is the order of the resonance.
Diophantine conditions describe a measure-theoretically large complement of a neighborhood of the
(dense!) set of all resonances.
Separatrices Consider a hyperbolic equilibrium, xed or
periodic point or invariant torus. If the stable and unstable manifolds of such hyperbolic elements are codimension one immersed manifolds, then they are called
separatrices, since they separate domains of phase
space, for instance, basins of attraction.
Singularity theory A function H : Rn ! R has a critical point z 2 Rn where DH(z) vanishes. In local coordinates we may arrange z D 0 (and similarly that it is mapped to zero as well). Two germs
K : (Rn ; 0) ! (R; 0) and N : (Rn ; 0) ! (R; 0) represent the same function H locally around z if and only
if there is a dieomorphism on Rn satisfying
N D K :
The corresponding equivalence class is called a singularity.
Structurally stable A system is structurally stable if it is
topologically equivalent to all nearby systems, where
nearby is measured in an appropriate topology on the
space of systems, like the Whitney Ck -topology [72].
A family is structurally stable if for every nearby family
there is a re-parametrization such that all corresponding systems are topologically equivalent.
Definition of the Subject
The fundamental problem of mechanics is to study Hamiltonian systems that are small perturbations of integrable
systems. Also, perturbations that destroy the Hamiltonian
character are important, be it to study the eect of a small
amount of friction, or to further the theory of dissipative systems themselves which surprisingly often revolves
around certain well-chosen Hamiltonian systems. Furthermore there are approaches like KAM theory that historically were rst applied to Hamiltonian systems. Typically perturbation theory explains only part of the dynamics, and in the resulting gaps the orderly unperturbed motion is replaced by random or chaotic motion.
Introduction
We outline perturbation theory from a general point of
view, illustrated by a few examples.
The Perturbation Problem

The aim of perturbation theory is to approximate a given
dynamical system by a more familiar one, regarding the
former as a perturbation of the latter. The problem then
is to deduce certain dynamical properties from the unperturbed to the perturbed case.
What is familiar may or may not be a matter of taste, at
least it depends a lot on the dynamical properties of ones
interest. Still the most frequently used unperturbed systems are:
Linear systems
Integrable Hamiltonian systems, compare with Dynamics of Hamiltonian Systems and references therein
Normal form truncations, compare with Normal
Forms in Perturbation Theory and references therein
Etc.
To some extent the second category can be seen as a special
case of the third. To avoid technicalities in this section we
assume all systems to be suciently smooth, say of class
C 1 or real analytic. Moreover in our considerations " will
be a real parameter. The unperturbed case always corresponds to " D 0 and the perturbed one to " 0 or " > 0.
Examples of Perturbation Problems To begin with
consider the autonomous dierential equation
x C "x C
dV
(x) D 0 ;
dx
modeling an oscillator with small damping. Rewriting this

equation of motion as a planar vector eld
x D
dV
(x) ;
dx
we consider the energy H(x; y) D 12 y 2 C V(x). For " D 0

the system is Hamiltonian with Hamiltonian function H.
Indeed, generally we have H(x;

y) D "y 2 , implying that
for " > 0 there is dissipation of energy. Evidently for " 0
the system is no longer Hamiltonian.
The reader is invited to compare the phase portraits of
the cases " D 0 and " > 0 for V (x) D cos x (the pendu1
lum) or V(x) D 12 x 2 C 24
bx 4 (Dung).
Another type of example is provided by the non-autonomous equation
x C
a vector eld, we obtain

t D 1
x D
y D
dV
(x) C " f (x; y; t) ;
dx
now on the generalized phase space R3 D ft; x; yg. In

the case where the t-dependence is periodic, we can take
S1 R2 for (generalized) phase space.
Remark
A small variation of the above driven system concerns
a parametrically forced oscillator like
x C (! 2 C " cos t) sin x D 0 ;
which happens to be entirely in the world of Hamiltonian systems.
It may be useful to study the Poincar or period mapping of such time periodic systems, which happens to
be a mapping of the plane. We recall that in the Hamiltonian cases this mapping preserves area. For general
reference in this direction see, e. g., [6,7,27,66].
There are lots of variations and generalizations. One example is the solar system, where the unperturbed case consists of a number of uncoupled two-body problems concerning the Sun and each of the planets, and where the interaction between the planets is considered as small [6,9,
107,108].
Remark
y D "y
dV
(x) D " f (x; x ; t) ;
dx
which can be regarded as the equation of motion of an

oscillator with small external forcing. Again rewriting as
One variation is a restriction to fewer bodies, for

example only three. Examples of this are systems
like SunJupiterSaturn, EarthMoonSun or Earth
MoonSatellite.
Often Sun, Moon and planets are considered as point
masses, in which case the dynamics usually are modeled as a Hamiltonian system. It is also possible to extend this approach taking tidal eects into account,
which have a non-conservative nature.
The Solar System is close to resonance, which makes
application of KAM theory problematic. There exist,
however, other integrable approximations that take
resonance into account [3,63].
Quite another perturbation setting is local, e. g., near an
equilibrium point. To x thoughts consider
x D Ax C f (x) ;
x 2 Rn
4517
4518
with A 2 gl(n; R), f (0) D 0 and Dx f (0) D 0. By the scaling x D "x we rewrite the system to
x D Ax C "g(x ) :
So, here we take the linear part as an unperturbed system. Observe that for small " the perturbation is small on
a compact neighborhood of x D 0.
This setting also has many variations. In fact, any
normal form approximation may be treated in this
way Normal Forms in Perturbation Theory. Then the
normalized truncation forms the unperturbed part and the
higher order terms the perturbation.
Remark In the above we took the classical viewpoint
which involves a perturbation parameter controlling the
size of the perturbation. Often one can generalize this by
considering a suitable topology (like the Whitney topologies) on the corresponding class of systems [72]. Also
compare with Normal Forms in Perturbation Theory,
KolmogorovArnoldMoser (KAM) Theory and Dynamics of Hamiltonian Systems.
Questions of Persistence
What are the kind of questions perturbation theory asks?
A large class of questions concerns the persistence of certain dynamical properties as known for the unperturbed
case. To x thoughts we give a few examples.
To begin with consider equilibria and periodic orbits.
So we put
x D f (x; ") ;
x 2 Rn ; " 2 R ;
(1)
for a map f : RnC1 ! Rn . Recall that equilibria are given

by the equation f (x; ") D 0. The following theorem that
continues equilibria in the unperturbed system for " 0,
is a direct consequence of the implicit function theorem.
Theorem 1 (Persistence of equilibria)
f (x0 ; 0) D 0 and that
Suppose that
Dx f (x0 ; 0) has maximal rank :

Then there exists a local arc " 7! x(") with x(0) D x0 such
that
f (x("); ") 0 :
Periodic orbits can be approximated in a similar way. Indeed, let the system (1) for D 0 have a periodic orbit 0 .
Let be a local transversal section of 0 and P0 : !
the corresponding Poincar map. Then P0 has a xed
point x0 2 \ 0 . By transversality, for j"j small, a local

Poincar map P" : ! is well-dened for (1). Observe
that xed points x" of P" correspond to periodic orbits "
of (1). We now have, again as another direct consequence
of the implicit function theorem.
Theorem 2 (Persistence of periodic orbits) In the above
assume that
P0 (x0 ) D x0 and Dx P0 (x0 ) has no eigenvalue 1 :
Then there exists a local arc " 7! x(") with x(0) D x0 such
that
P" (x(")) x" :
Remark
Often the conditions of Theorem 2 are not easy to verify. Sometimes it is useful here to use Floquet Theory,
see [97]. In fact, if T 0 is the period of 0 and 0 its
Floquet matrix, then Dx P0 (x0 ) D exp(T0 0 ).
The format of the Theorems 1 and 2 with the perturbation parameter " directly allows for algorithmic approaches. One way to proceed is by perturbation series,
leading to asymptotic formulae that in the real analytic
setting have positive radius of convergence. In the latter
case the names of Poincar and Lindstedt are associated
with the method, cf. [10].
Also numerical continuation programmes exist based
on the Newton method.
The Theorems 1 and 2 can be seen as special cases of
a a general theorem for normally hyperbolic invariant
manifolds [73], Theorem 4.1. In all cases a contraction
principle on a suitable Banach space of graphs leads to
persistence of the invariant dynamical object.
This method in particular yields existence and persistence of stable and unstable manifolds [53,54].
Another type of dynamics subject to perturbation theory
is quasi-periodic. We emphasize that persistence of (Diophantine) quasi-periodic invariant tori occurs both in the
conservative setting and in many others, like in the reversible and the general (dissipative) setting. In the latter
case this leads to persistent occurrence of families of quasiperiodic attractors [125]. These results are in the domain
of KolmogorovArnoldMoser (KAM) theory. For details
we refer to Sect. KAM Theory: An Overview below or
to [24], KolmogorovArnoldMoser (KAM) Theory,
the former reference containing more than 400 references
in this area.
Remark
Concerning the Solar System, KAM theory always has
aimed at proving that it contains many quasi-periodic
motions, in the sense of positive Liouville measure.
This would imply that there is positive probability that
a given initial condition lies on such a stable quasi-periodic motion [3,63], however, also see [85].
Another type of result in this direction compares the
distance of certain individual solutions of the perturbed
and the unperturbed system, with coinciding initial
conditions over time scales that are long in terms of ".
Compare with [24].
Apart from persistence properties related to invariant
manifolds or individual solutions, the aim can also be
to obtain a more global persistence result. As an example of this we mention the HartmanGrobman Theorem,
e. g., [7,116,123]. Here the setting once more is
x D Ax C f (x) ;
x 2 Rn ;
with A 2 gl(n; R), f (0) D 0 and Dx f (0) D 0. Now we assume A to be hyperbolic (i. e., with no purely imaginary
eigenvalues). In that case the full system, near the origin,
is topologically conjugated to the linear system x D Ax.
Therefore all global, qualitative properties of the unperturbed (linear) system are persistent under perturbation to
the full system. For details on these notions see the above
references, also compare with, e. g., [30].
It is said that the hyperbolic linear system x D Ax is
(locally) structurally stable. This kind of thinking was introduced to the dynamical systems area by Thom [133],
with a rst, successful application to catastrophe theory.
For further details, see [7,30,69,116].
General Dynamics
We give a few remarks on the general dynamics in a neighborhood of Hamiltonian KAM tori. In particular this concerns so-called superexponential stickiness of the KAM tori
and adiabatic stability of the action variables, involving the
so-called Nekhoroshev estimate.
To begin with, emphasize the following dierence between the cases n D 2 and n 3 in the classical KAM theorem of Subsect. Classical KAM Theory. For n D 2 the
level surfaces of the Hamiltonian are three-dimensional,
while the Lagrangian tori have dimension two and hence
codimension one in the energy hypersurfaces. This means
that for open sets of initial conditions, the evolution curves
are forever trapped in between KAM tori, as these tori foliate over nowhere dense sets of positive measure. This im-
plies perpetual adiabatic stability of the action variables.

In contrast, for n 3 the Lagrangian tori have codimension n 1 > 1 in the energy hypersurfaces and evolution
curves may escape.
This actually occurs in the case of so-called Arnold
diusion. The literature on this subject is immense, and
we here just quote [5,9,93,109], for many more references
see [24].
Next we consider the motion in a neighborhood of
the KAM tori, in the case where the systems are real analytic or at least Gevrey smooth. For a denition of Gevrey
regularity see [136]. First we mention that, measured in
terms of the distance to the KAM torus, nearby evolution
curves generically stay nearby over a superexponentially
long time [102,103]. This property often is referred to as
superexponential stickiness of the KAM tori, see [24] for
more references.
Second, nearly integrable Hamiltonian systems, in
terms of the perturbation size, generically exhibit exponentially long adiabatic stability of the action variables, see
e. g. [15,88,89,90,93,103,109,110,113,120], Nekhoroshev
Theory and many others, for more references see [24].
This property is referred to as the Nekhoroshev estimate
or the Nekhoroshev theorem. For related work on perturbations of so-called superintegrable systems, also see [24]
and references therein.
Chaos
In the previous subsection we discussed persistent and
some non-persistent features of dynamical systems under small perturbations. Here we discuss properties related to splitting of separatrices, caused by generic perturbations.
A rst example was met earlier, when comparing the
pendulum with and without (small) damping. The unperturbed system is the undamped one and this is a Hamiltonian system. The perturbation however no longer is
Hamiltonian. We see that the equilibria are persistent, as
they should be according to Theorem 1, but that none of
the periodic orbits survives the perturbation. Such qualitative changes go with perturbing away from the Hamiltonian setting.
Similar examples concern the breaking of a certain
symmetry by the perturbation. The latter often occurs in
the case of normal form approximations. Then the normalized truncation is viewed as the unperturbed system,
which is perturbed by the higher order terms. The truncation often displays a reasonable amount of symmetry (e. g.,
toroidal symmetry), which generically is forbidden for the
class of systems under consideration, e. g. see [25].
4519
4520
Hamiltonian Perturbation Theory (and Transition to Chaos), Figure 1

Chaos in the parametrically forced pendulum. Left: Poincar map P!;" near the 1 : 2 resonance ! D
Right: A dissipative analogue
To x thoughts we reconsider the conservative example

x C (! 2 C " cos t) sin x D 0
of the previous section. The corresponding (time dependent, Hamiltonian [6]) vector eld reads
t D
x D
y D (! 2 C " cos t) sin x :
and for " > 0 not too small.
more sophisticated tools see [65]. Often this leads to elliptic (Abelian) integrals.
In nearly integrable systems chaos can occur. This
fact is at the heart of the celebrated non-integrability of
the three-body problem as addressed by Poincar [12,59,
107,108,118]. A long standing open conjecture is that the
clouds of points as visible in Fig. 1, left, densely ll sets of
positive area, thereby leading to ergodicity [9].
In the case of dissipation, see Fig. 1, right, we conjecture the occurrence of a Hnon-like strange attractor [14,22,126].
Let P!;" : R2 ! R2 be the corresponding (area-preserving) Poincar map. Let us consider the unperturbed map
P!;0 which is just the ow over time 2 of the free
pendulum x C ! 2 sin x D 0. Such a map is called integrable, since it is the stroboscopic map of a two-dimensional vector eld, hence displaying the R-symmetry of
a ow. When perturbed to the nearly integrable case " 0,
this symmetry generically is broken. We list a few of the
generic properties for such maps [123]:
Remark
The homoclinic and heteroclinic points occur at

transversal intersections of the corresponding stable
and unstable manifolds.
The periodic points of period less than a given bound
are isolated.
One Degree of Freedom
This means generically that the separatrices split and that

the resonant invariant circles lled with periodic points
with the same (rational) rotation number fall apart. In
any concrete example the issue remains whether or not
it satises appropriate genericity conditions. One method
to check this is due to Melnikov, compare [66,137], for
1
2
The persistent occurrence of periodic points of a given

rotation number follows from the PoincarBirkho
xed point theorem [74,96,107], i. e., on topological
grounds.
The above arguments are not restricted to the conservative setting, although quite a number of unperturbed
systems come from this world. Again see Fig. 1.
Planar Hamiltonian systems are always integrable and the

orbits are given by the level sets of the Hamiltonian function. This still leaves room for a perturbation theory. The
recurrent dynamics consists of periodic orbits, equilibria
and asymptotic trajectories forming the (un)stable manifolds of unstable equilibria. The equilibria organize the
phase portrait, and generically all equilibria are elliptic
(purely imaginary eigenvalues) or hyperbolic (real eigenvalues), i. e. there is no equilibrium with a vanishing eigenvalue. If the system depends on a parameter such van-
ishing eigenvalues may be unavoidable and it becomes

possible that the corresponding dynamics persist under
perturbations.
Perturbations may also destroy the Hamiltonian character of the ow. This happens especially where the starting point is a dissipative planar system and e. g. a scaling
leads for " D 0 to a limiting Hamiltonian ow. The perturbation problem then becomes twofold. Equilibria still
persist by Theorem 1 and hyperbolic equilibria moreover
persist as such, with the sum of eigenvalues of order O(").
Also for elliptic eigenvalues the sum of eigenvalues is of
order O(") after the perturbation, but here this number
measures the dissipation whence the equilibrium becomes
(weakly) attractive for negative values and (weakly) unstable for positive values. The one-parameter families of periodic orbits of a Hamiltonian system do not persist under
dissipative perturbations, the very fact that they form families imposes the corresponding xed point of the Poincar
mapping to have an eigenvalue one and Theorem 2 does
not apply. Typically only nitely many periodic orbits survive a dissipative perturbation and it is already a dicult
task to determine their number.
Hamiltonian Perturbations
The Dung oscillator has the Hamiltonian function
H(x; y) D
1 2
1
1
y C bx 4 C x 2
2
24
2
(2)
where b is a constant distinguishing the two cases b D 1

and is a parameter. Under variation of the parameter the
equations of motion
1
y D bx 3 x
6
H; (x; y) D
1 2
1
1
y C bx 4 C x 2 C x
2
24
2
of Hamiltonian systems is indeed structurally stable. This

implies not only that all equilibria of a Hamiltonian perturbation of the Dung oscillator have a local ow equivalent to the local ow near a suitable equilibrium in this
two-parameter family, but that every one-parameter family of Z2 -symmetric Hamiltonian systems that is a perturbation of (2) has equivalent dynamics. For more details
see [36] and references therein.
This approach applies mutatis mutandis to every nondegenerate planar singularity, cf. [69,130]. At an equilibrium all partial derivatives of the Hamiltonian vanish and
the resulting singularity is called non-degenerate if it has
nite multiplicity, which implies that it admits a versal unfolding H with nitely many parameters. The family of
Hamiltonian systems dened by this versal unfolding contains all possible (local) dynamics that the initial equilibrium may be perturbed to. Imposing additional discrete
symmetries is immediate, the necessary symmetric versal
unfolding is obtained by averaging
HG D
1 X
H g
jGj
along the orbits of the symmetry group G.
display a Hamiltonian pitchfork bifurcation, supercritical

for positive b and subcritical in case b is negative. Correspondingly, the linearization at the equilibrium x D 0 of
the anharmonic oscillator D 0 is given by the matrix
0
0
where a; b; c 2 R are nonzero constants, for instance

a D b D c D 1. Note that this is a completely dierent unfolding of the parabolic equilibrium at the origin. A closer
look at the phase portraits and in particular at the Hamiltonian function of the Hamiltonian pitchfork bifurcation
reveals the symmetry x 7! x of the Dung oscillator.
This suggests the addition of the non-symmetric term x.
The resulting two-parameter family
g2G
x D y
1
0
whence this equilibrium is parabolic.

The typical way in which a parabolic equilibrium bifurcates is the center-saddle bifurcation. Here the Hamiltonian reads
H(x; y) D
1 2 1 3
ay C bx C c x
2
6
(3)
Dissipative Perturbations
In a generic dissipative system all equilibria are hyperbolic.
Qualitatively, i. e. up to topological equivalence, the local dynamics is completely determined by the number of
eigenvalues with positive real part. Those hyperbolic equilibria that can appear in Hamiltonian systems (the eigenvalues forming pairs ) do not play an important role.
Rather, planar Hamiltonian systems become important as
a tool to understand certain bifurcations triggered o by
non-hyperbolic equilibria. Again this requires the system
to depend on external parameters.
The simplest example is the Hopf bifurcation, a co-dimension one bifurcation where an equilibrium loses stability as the pair of eigenvalues crosses the imaginary axis,
4521
4522
say at i. At the bifurcation the linearization is a Hamiltonian system with an elliptic equilibrium (the co-dimension one bifurcations where a single eigenvalue crosses the
imaginary axis through 0 do not have a Hamiltonian linearization). This limiting Hamiltonian system has a oneparameter family of periodic orbits around the equilibrium, and the non-linear terms determine the fate of these
periodic orbits. The normal form of order three reads

x D y 1 C b(x C y ) C x C a(x 2 C y 2 )

x D x 1 C b(x 2 C y 2 ) C y C a(x 2 C y 2 )
2
and is Hamiltonian if and only if ( ; a) D (0; 0). The sign

of the coecient distinguishes between the supercritical
case a > 0, in which there are no periodic orbits coexisting
with the attractive equilibria (i. e. when < 0) and one attracting periodic orbit for each > 0 (coexisting with the
unstable equilibrium), and the subcritical case a < 0, in
which the family of periodic orbits is unstable and coexists
with the attractive equilibria (with no periodic orbits for
parameters > 0). As ! 0 the family of periodic orbits
shrinks down to the origin, so also this Hamiltonian feature is preserved.
Equilibria with a double eigenvalue 0 need two parameters to persistently occur in families of dissipative systems. The generic case is the TakensBogdanov bifurcation. Here the linear part is too degenerate to be helpful,
but the nonlinear Hamiltonian system dened by (3) with
a D 1 D c and b D 3 provides the periodic and heteroclinic orbit(s) that constitute the nontrivial part of the bifurcation diagram. Where discrete symmetries are present,
e. g. for equilibria in dissipative systems originating from
other generic bifurcations, the limiting Hamiltonian system exhibits that same discrete symmetry. For more details
see [54,66,82] and references therein.
The continuation of certain periodic orbits from an
unperturbed Hamiltonian system under dissipative perturbation can be based on Melnikov-like methods, again
see [66,137]. As above, this often leads to Abelian integrals,
for instance to count the number of periodic orbits that
branch o.
Reversible Perturbations
A dynamical system that admits a reection symmetry R
mapping trajectories '(t; z0 ) to trajectories '(t; R(z0 )) is
called reversible. In the planar case we may restrict to the
reversing reection
R:
R2
(x; y)
!
7!
R2
(x; y) :
(4)
All Hamiltonian functions H D 12 y 2 C V(x) which have

an interpretation kinetic C potential energy are reversible, and in general the class of reversible systems is
positioned between the class of Hamiltonian systems and
the class of dissipative systems. A guiding example is the
perturbed Dung oscillator (with the roles of x and y exchanged so that (4) remains the reversing symmetry)
1
x D y 3 y C "x y
6
y D x
that combines the Hamiltonian character of the equilibrium at the origin with the dissipative character of the two
other equilibria. Note that all orbits outside the homoclinic
loop are periodic.
There are two ways in which the reversing symmetry (4) imposes a Hamiltonian character on the dynamics.
An equilibrium that lies on the symmetry line fy D 0g has
a linearization that is itself a reversible system and consequently the eigenvalues are subject to the same constraints
as in the Hamiltonian case. (For equilibria z0 that do not lie
on the symmetry line the reection R(z0 ) is also an equilibrium, and it is to the union of their eigenvalues that
these constraints still apply.) Furthermore, every orbit that
crosses fy D 0g more than once is automatically periodic,
and these periodic orbits form one-parameter families. In
particular, elliptic equilibria are still surrounded by periodic orbits.
The dissipative character of a reversible system is most
obvious for orbits that do not cross the symmetry line.
Here R merely maps the orbit to a reected counterpart.
The above perturbed Dung oscillator exemplies that
the character of an orbit crossing fy D 0g exactly once is
undetermined. While the homoclinic orbit of the saddle
at the origin has a Hamiltonian character, the heteroclinic
orbits between the other two equilibria behave like in a dissipative system.
Perturbations of Periodic Orbits
The perturbation of a one-degree-of-freedom system by
a periodic forcing is a perturbation that changes the phase
space. Treating the time variable t as a phase space variable leads to the extended phase space S1 R2 and equilibria of the unperturbed system become periodic orbits,
inheriting the normal behavior. Furthermore introducing
an action conjugate to the angle t yields a Hamiltonian
system in two degrees of freedom.
While the one-parameter families of periodic orbits
merely provide the typical recurrent motion in one degree
of freedom, they form special solutions in two or more de-
grees of freedom. Arcs of elliptic periodic orbits are particularly instructive. Note that these occur generically in both
the Hamiltonian and the reversible context.
Conservative Perturbations
Along the family of elliptic periodic orbits a pair ei of
Floquet multipliers passes regularly through roots of unity.
Generically this happens on a dense set of parameter values, but for xed denominator q in ei D e2 ip/q the
corresponding energy values are isolated. The most important of such resonances are those with small denominators q.
For q D 1 generically a periodic center-saddle bifurcation takes place where an elliptic and a hyperbolic periodic
orbit meet at a parabolic periodic orbit. No periodic orbit
remains under further variation of a suitable parameter.
The generic bifurcation for q D 2 is the period-doubling bifurcation where an elliptic periodic orbit turns hyperbolic (or vice versa) when passing through a parabolic
periodic orbit with Floquet multipliers 1. Furthermore,
a family of periodic orbits with twice the period emerges
from the parabolic periodic orbit, inheriting the normal
linear behavior from the initial periodic orbit.
In case q D 3, and possibly also for q D 4, generically
two arcs of hyperbolic periodic orbits emerge, both with
three (resp. four) times the period. One of these extends
for lower and the other for higher parameter values. The
initial elliptic periodic orbit momentarily loses its stability
due to these approaching unstable orbits.
Denominators q 5 (and also the second possibility
for q D 4) lead to a pair of subharmonic periodic orbits
of q times the period emerging either for lower or for
higher parameter values. This is (especially for large q)
comparable to the behavior at Diophantine ei where
a family of invariant tori emerges, cf. Sect. Invariant
Curves of Planar Dieomorphisms below.
For a single pair ei of Floquet multipliers this
behavior is traditionally studied for the (iso-energetic)
Poincar-mapping, cf. [92] and references therein. However, the above description remains true in higher dimensions, where additionally multiple pairs of Floquet multipliers may interact. An instructive example is the Lagrange top, the sleeping motion of which is gyroscopically
stabilized after a periodic Hamiltonian Hopf bifurcation;
see [56] for more details.
Dissipative Perturbations
There exists a large class of local bifurcations in the dissipative setting, that can be arranged in a perturbation
theory setting, where the unperturbed system is Hamil-
tonian. The arrangement consists of changes of variables and rescaling. An early example of this is the BogdanovTakens bifurcation [131,132]. For other examples
regarding nilpotent singularities, see [23,40] and references therein.
To x thoughts, consider families of planar maps and
let the unperturbed Hamiltonian part contain a center
(possibly surrounded by a homoclinic loop). The question
then is which of these persist when adding the dissipative
perturbation.
Usually only a denite nite number persists. As in
Subsect. Chaos, a Melnikov function can be invoked
here, possibly again leading to elliptic (Abelian) integrals,
Picard Fuchs equations, etc. For details see [61,124] and
references therein.
Invariant Curves of Planar Diffeomorphisms
This section starts with general considerations on circle dieomorphisms, in particular focusing on persistence
properties of quasi-periodic dynamics. Our main references are [2,24,29,31,70,71,139,140]. For a denition of rotation number, see [58]. After this we turn to area preserving maps of an annulus where we discuss Mosers twist
map theorem [104], also see [24,29,31]. The section is concluded by a description of the holomorphic linearization of
a xed point in a planar map [7,101,141,142].
Our main perspective will be perturbative, where we
consider circle maps near a rigid rotation. It turns out that
generally parameters are needed for persistence of quasiperiodicity under perturbations. In the area preserving setting we consider perturbations of a pure twist map.
Circle Maps
We start with the following general problem. Given a twoparameter family
P;" : T 1 ! T 1 ;
x 7! x C 2 C "a(x; ; ")
of circle maps of class C 1 . It turns out to be convenient to

view this two-parameter family as a one-parameter family
of maps
P" : T 1 [0; 1] ! T 1 [0; 1];
(x; ) 7! (x C 2 C "a(x; ; "); )
of the cylinder. Note that the unperturbed system P0 is
a family of rigid circle rotations, viewed as a cylinder map,
where the individual map P;0 has rotation number . The
question now is what will be the fate of this rigid dynamics
for 0 j"j
1.
4523
4524
The classical way to address this question is to look for

a conjugation " , that makes the following diagram commute
T 1 [0; 1]
" "
!
P"
T 1 [0; 1]
" "
T 1 [0; 1]
!
P0
T 1 [0; 1] ;
i. e., such that

P" " D " P0 :
Due to the format of P" we take " as a skew map
" (x; ) D (x C "U(x; ; "); C "(; ")) ;
which leads to the nonlinear equation
U(x C 2 ; ; ") U(x; ; ")
D 2(; ") C a (x C "U(x; ; "); C "(; "); ")
in the unknown maps U and . Expanding in powers of "
and comparing at lowest order yields the linear equation
Hamiltonian Perturbation Theory (and Transition to Chaos),

Figure 2
Skew cylinder map, conjugating (Diophantine) quasi-periodic invariant circles of P0 and P"
U0 (x C 2 ; ) U0 (x; ) D 20 () C a0 (x; )

which can be directly solved by Fourier series. Indeed,
writing
X
a0k ()eikx ;
a0 (x; ) D
k2Z
U0 (x; ) D
U0k ()eikx
k2Z
we nd 0 D 1/(2)a00 and
U0k () D
a0k ()
:
e2 ik 1
It follows that in general a formal solution exists if and

only if 2 R n Q. Still, the accumulation of e2 ik 1
on 0 leads to the celebrated small divisors [9,108], also
see [24,29,31,55].
The classical solution considers the following Diophantine non-resonance conditions. Fixing > 2 and
> 0 consider 2 [0; 1] such that for all rationals p/q
p q :
(5)
Theorem 3 (Circle Map Theorem) For suciently

small and for the perturbation "a suciently small
in the C 1 -topology, there exists a C 1 transformation
" : T 1 [0; 1] ! T 1 [0; 1], conjugating the restriction
P0 j[0;1]; to a subsystem of P" .
Theorem 3 in the present structural stability formulation (compare with Fig. 2) is a special case of the results
in [29,31]. We here speak of quasi-periodic stability. For
earlier versions see [2,9].
Remark
Rotation numbers are preserved by the map " and irrational rotation numbers correspond to quasi-periodicity. Theorem 3 thus ensures that typically quasi-periodicity occurs with positive measure in the parameter
space. Note that since Cantor sets are perfect, quasi-periodicity typically has a non-isolated occurrence.
The map " has no dynamical meaning inside the gaps.
The gap dynamics in the case of circle maps can be illustrated by the Arnold family of circle maps [2,7,58],
given by
P;" (x) D x C 2 C " sin x
This subset of such s is denoted by [0; 1]; and is wellknown to be nowhere dense but of large measure as > 0
gets small [115]. Note that Diophantine numbers are
irrational.
which exhibits a countable union of open resonance

tongues where the dynamics is periodic, see Fig. 3. Note
that this map only is a dieomorphism for j"j < 1.
Remark

Figure 3
Arnold resonance tongues; for " 1 the maps are endomorphic
We like to mention that non-perturbative versions of

Theorem 3 have been proven in [70,71,139].
For simplicity we formulated Theorem 3 under
C 1 -regularity, noting that there exist many ways to
generalize this. On the one hand there exist Ck -versions
for nite k and on the other hand there exist ne tunings in terms of real-analytic and Gevrey regularity. For
details we refer to [24,31] and references therein. This
same remark applies to other results in this section and
in Sect. KAM Theory: An Overview on KAM theory.
A possible application of Theorem 3 runs as follows. Consider a system of weakly coupled Van der Pol oscillators
y1 C c1 y1 C a1 y1 C f1 (y1 ; y1 ) D "g1 (y1 ; y2 ; y1 ; y2 )
y2 C c2 y2 C a2 y2 C f2 (y2 ; y2 ) D "g2 (y1 ; y2 ; y1 ; y2 ) :
Writing y j D z j ; j D 1; 2, one obtains a vector eld in
the four-dimensional phase space R2 R2 D f(y1 ; z1 );
(y2 ; z2 )g. For " D 0 this vector eld has an invariant
two-torus, which is the product of the periodic motions
of the individual Van der Pol oscillations. This twotorus is normally hyperbolic and therefore persistent for
j"j
1 [73]. In fact the torus is an attractor and we can dene a Poincar return map within this torus attractor. If we
include some of the coecients of the equations as parameters, Theorem 3 is directly applicable. The above statements on quasi-periodic circle maps then directly translate
to the case of quasi-periodic invariant two-tori. Concerning the resonant cases, generically a tongue structure like
in Fig. 3 occurs; for the dynamics corresponding to parameter values inside such a tongue one speaks of phase lock.
The celebrated synchronization of Huygens clocks [77]

is related to a 1 : 1 resonance, meaning that the corresponding Poincar map would have its parameters
in the main tongue with rotation number 0. Compare
with Fig. 3.
There exist direct generalizations to cases with n-oscillators (n 2 N), leading to families of invariant n-tori
carrying quasi-periodic ow, forming a nowhere dense
set of positive measure. An alteration with resonance
occurs as roughly sketched in Fig. 3. In higher dimension the gap dynamics, apart from periodicity, also can
contain strange attractors [112,126]. We shall come
back to this subject in a later section.
Area-Preserving Maps
The above setting historically was preceded by an area preserving analogue [104] that has its origin in the Hamiltonian dynamics of frictionless mechanics.
Let R2 n f(0; 0)g be an annulus, with sympectic
polar coordinates ('; I) 2 T 1 K, where K is an interval.
Moreover, let D d' ^ dI be the area form on .
We consider a -preserving smooth map P" : !
of the form
P" ('; I) D (' C 2 (I); I) C O(") ;
where we assume that the map I 7! (I) is a (local) dieomorphism. This assumption is known as the twist condition and P" is called a twist map. For the unperturbed case
" D 0 we are dealing with a pure twist map and its dynamics are comparable to the unperturbed family of cylinder
maps as met in Subsect. Circle Maps. Indeed it is again
a family of rigid rotations, parametrized by I and where
P0 (:; I) has rotation number (I). In this case the question is what will be the fate of this family of invariant circles, as well as with the corresponding rigidly rotational
dynamics.
Regarding the rotation number we again introduce
Diophantine conditions. Indeed, for > 2 and > 0 the
subset [0; 1]; is dened as in (5), i. e., it contains all
2 [0; 1], such that for all rationals p/q
p q :
q
Pulling back [0; 1]; along the map we obtain a subset
; .
Theorem 4 (Twist Map Theorem [104]) For suciently small, and for the perturbation O(") suciently
4525
4526
small in C 1 -topology, there exists a C 1 transformation

" : ! , conjugating the restriction P0 j#; to a subsystem of P" .
As in the case of Theorem 3 again we chose the formulation of [29,31]. Largely the remarks following Theorem 3
also apply here.
Remark
Compare the format of the Theorems 3 and 4 and observe that in the latter case the role of the parameter
has been taken by the action variable I. Theorem 4 implies that typically quasi-periodicity occurs with positive measure in phase space.
In the gaps typically we have coexistence of periodicity,
quasi-periodicity and chaos [6,9,35,107,108,123,137].
The latter follows from transversality of homo- and
heteroclinic connections that give rise to positive topological entropy. Open problems are whether the corresponding Lyapunov exponents also are positive, compare with the discussion at the end of the introduction.
Similar to the applications of Theorem 3 given at the end
of Subsect. Circle Maps, here direct applications are possible in the conservative setting. Indeed, consider a system
of weakly coupled pendula
@U
(y1 ; y2 )
@y1
@U
y2 C 22 sin y2 D "
(y1 ; y2 ) :
@y2
y1 C 12 sin y1 D "
Writing y j D z j , j D 1; 2 as before, we again get a vector eld in the four-dimensional phase space R2 R2 D
f(y1 ; y2 ); (z1 ; z2 )g. In this case the energy
H" (y1 ; y2 ; z1 ; z2 )
1
1
D z12 C z22 12 cos y1 22 cos y2 C "U(y1 ; y2 )
2
2
is a constant of motion. Restricting to a three-dimensional
energy surface H"1 D const:, the iso-energetic Poincar
map P" is a twist map and application of Theorem 4 yields
the conclusion of quasi-periodicity (on invariant two-tori)
occurring with positive measure in the energy surfaces
of H" .
Remark As in the dissipative case this example directly
generalizes to cases with n oscillators (n 2 N), again leading to invariant n-tori with quasi-periodic ow. We shall
return to this subject in a later section.
Linearization of Complex Maps

The Subsects. Circle Maps and Area-Preserving Maps
both deal with smooth circle maps that are conjugated to
rigid rotations. Presently the concern is with planar holomorphic maps that are conjugated to a rigid rotation on an
open subset of the plane. Historically this is the rst time
that a small divisor problem was solved [7,101,141,142]
and Perturbative Expansions, Convergence of.
Complex Linearization Given is a holomorphic germ
F : (C; 0) ! (C; 0) of the form F(z) D z C f (z), with
f (0) D f 0 (0) D 0. The problem is to nd a biholomorphic
germ : (C; 0) ! (C; 0) such that
F D :
Such a dieomorphism is called a linearization of F
near 0.
We begin with the formal approach. Given the series
P
P
f (z) D j2 f j z j , we look for (z) D z C j2 j z j . It
turns out that a solution always exists whenever 0 is
not a root of unity. Indeed, direct computation reveals the
following set of equations that can be solved recursively:
For j D 2 get the equation (1 )2 D f2
For j D 3 get the equation (1 2 )3 D f3 C 2 f2 2
For j D n get the equation (1 n1 ) n D f n Cknown.
The question now reduces to whether this formal solution
has a positive radius of convergence.
The hyperbolic case 0 < j j 1 was already solved by
Poincar, for a description see [7]. The elliptic case j j D 1
again has small divisors and was solved by Siegel when for
some > 0 and > 2 we have the Diophantine non-resonance condition
p
j e
2 i q
j jqj :
The corresponding set of constitutes a set of full measure

in T 1 D f g.
Yoccoz [141] completely solved the elliptic case using
the Bruno-condition. If
D e2 i and when
pn
qn
is the nth convergent in the continued fraction expansion

of then the Bruno-condition reads
X log(q nC1 )
n
qn
<1:
This condition turns out to be necessary and sucient

for having positive radius of convergence [141,142].
Cremers Example in Hermans Version

consider the map
As an example
F(z) D z C z ;
where 2 T 1 is not a root of unity.
Observe that a point z 2 C is a periodic point of F with
period q if and only if F q (z) D z, where obviously
q
F q (z) D q z C C z2 :
Writing

q
F q (z) z D z q 1 C C z2 1 ;
the period q periodic points exactly are the roots of the
right hand side polynomial. Abbreviating N D 2q 1, it
directly follows that, if z1 ; z2 ; : : : ; z N are the nontrivial
roots, then for their product we have
z1 z2 : : : z N D q 1 :
It follows that there exists a nontrivial root within radius
j q 1j1/N
of z D 0.
Now consider the set of T 1 dened as follows:
2 whenever
lim inf j q 1j1/N D 0 :
q!1
It can be directly shown that is residual, again compare

with [115]. It also follows that for 2 linearization is
impossible. Indeed, since the rotation is irrational, the existence of periodic points in any neighborhood of z D 0
implies zero radius of convergence.
Remark
Notice that the residual set is in the complement of
the full measure set of all Diophantine numbers, again
see [115].
Considering 2 T 1 as a parameter, we see a certain
analogy of these results on complex linearization with
the Theorems 3 and 4. Indeed, in this case for a full
measure set of s on a neighborhood of z D 0 the map
F D F is conjugated to a rigid irrational rotation.
Such a domain in the z-plane often is referred to as
a Siegel disc. For a more general discussion of these and
of Herman rings, see [101].
KAM Theory: An Overview

In Sect. Invariant Curves of Planar Dieomorphisms we
described the persistent occurrence of quasi-periodicity in
the setting of dieomorphisms of the circle or the plane.
The general perturbation theory of quasi-periodic motions
is known under the name KolmogorovArnoldMoser (or
KAM) theory and discussed extensively elsewhere in this
encyclopedia KolmogorovArnoldMoser (KAM) Theory. Presently we briey summarize parts of this KAM
theory in broad terms, as this ts in our considerations,
thereby largely referring to [4,80,81,119,121,143,144], also
see [20,24,55].
In general quasi-periodicity is dened by a smooth
conjugation. First on the n-torus T n D Rn /(2 Z)n consider the vector eld
X! D
n
X
jD1
!j
@
;
@' j
where !1 ; !2 ; : : : ; !n are called frequencies [43,106].

Now, given a smooth (say, of class C 1 ) vector eld X
on a manifold M, with T M an invariant n-torus,
we say that the restriction Xj T is parallel if there exists
! 2 Rn and a smooth dieomorphism : T ! T n ,
such that (Xj T ) D X! . We say that Xj T is quasi-periodic if the frequencies !1 ; !2 ; : : : ; !n are independent
over Q.
A quasi-periodic vector eld Xj T leads to an integer
ane structure on the torus T. In fact, since each orbit is
dense, it follows that the self conjugations of X! exactly
are the translations of T n , which completely determine
the ane structure of T n . Then, given : T ! T n with
(Xj T ) D X! , it follows that the self conjugations of Xj T
determines a natural ane structure on the torus T. Note
that the conjugation is unique modulo translations in T
and T n .
Note that the composition of by a translation of
T n does not change the frequency vector !. However,
the composition by a linear invertible map S 2 GL(n; Z)
yields S X! D XS! . We here speak of an integer ane
structure [43].
Remark
The transition maps of an integer ane structure are
translations and elements of GL(n; Z).
The current construction is compatible with the integrable ane structure on the Liouville tori of an integrable Hamiltonian system [6]. Note that in that case
the structure extends to all parallel tori.
4527
4528
Classical KAM Theory

The classical KAM theory deals with smooth, nearly integrable Hamiltonian systems of the form
' D !(I) C " f (I; '; ")
I D "g(I; '; ") ;
(6)
where I varies over an open subset of Rn and ' over the

standard torus T n . Note that for " D 0 the phase space as
an open subset of Rn T n is foliated by invariant tori,
parametrized by I. Each of the tori is parametrized by '
and the corresponding motion is parallel (or multi- periodic or conditionally periodic) with frequency vector !(I).
Perturbation theory asks for persistence of the invariant n-tori and the parallelity of their motion for small
values of j"j. The answer that KAM theory gives needs
two essential ingredients. The rst ingredient is that of
Kolmogorov non-degeneracy which states that the map
I 2 Rn 7! !(I) 2 Rn is a (local) dieomorphism. Compare with the twist condition of Sect. Invariant Curves of
Planar Dieomorphisms. The second ingredient generalizes the Diophantine conditions (5) of that section as follows: for > n 1 and > 0 consider the set
n
R;
D f! 2 Rn j jh!; kij jkj ; k 2 Z n nf0gg: (7)
The following properties are more or less direct. First

n
R;
has a closed half line geometry in the sense that
n
n . Moreover,
and s 1 then also s! 2 R;
if ! 2 R;
n1
n
the intersection S
\ R; is a Cantor set of measure
n D O( ) as # 0, see Fig. 4.
S n1 n R;
Completely in the spirit of Theorem 4, the classical KAM theorem roughly states that a Kolmogorov nondegenerate nearly integrable system (6)" , for j"j
1 is
smoothly conjugated to the unperturbed version (6)0 , provided that the frequency map ! is co-restricted to the Dion . In this formulation smoothness has to
phantine set R;
be taken in the sense of Whitney [119,136], also compare
with [20,24,29,31,55,121].
As a consequence we may say that in Hamiltonian systems of n degrees of freedom typically quasi-periodic invariant (Lagrangian) n-tori occur with positive measure in
phase space. It should be said that also an iso-energetic
version of this classical result exists, implying a similar
conclusion restricted to energy hypersurfaces [6,9,21,24].
The Twist Map Theorem 4 is closely related to the iso-energetic KAM Theorem.
Remark
We chose the quasi-periodic stability format as in
Sect. Invariant Curves of Planar Dieomorphisms.

Figure 4
The Diophantine set Rn; has the closed half line geometry
and the intersection Sn1 \ Rn; is a Cantor set of measure
Sn1 n Rn; D O() as # 0
For regularity issues compare with a remark following

Theorem 3.
For applications we largely refer to the introduction
and to [24,31] and references therein.
Continuing the discussion on ane structures at the
beginning of this section, we mention that by means
of the symplectic form, the domain of the I-variables in
Rn inherits an ane structure [60], also see [91] and
references therein.
Statistical Mechanics deals with particle systems that
are large, often innitely large. The Ergodic Hypothesis
roughly says that in a bounded energy hypersurface, the
dynamics are ergodic, meaning that any evolution in the
energy level set comes near every point of this set.
The taking of limits as the number of particles tends
to innity is a notoriously dicult subject. Here we discuss a few direct consequences of classical KAM theory
for many degrees of freedom. This discussion starts with
Kolmogorovs papers [80,81], which we now present in
a slightly rephrased form. First, we recall that for Hamiltonian systems (say, with n degrees of freedom), typically
the union of Diophantine quasi-periodic Lagrangian invariant n-tori lls up positive measure in the phase space
and also in the energy hypersurfaces. Second, such a collection of KAM tori immediately gives rise to non-ergodic-
ity, since it clearly implies the existence of distinct invariant sets of positive measure. For background on Ergodic
Theory, see e. g. [9,27] and [24] for more references. Apparently the KAM tori form an obstruction to ergodicity,
and a question is how bad this obstruction is as n ! 1.
Results in [5,78] indicate that this KAM theory obstruction
is not too bad as the size of the system tends to innity.
In general the role of the Ergodic Hypothesis in Statistical
Mechanics has turned out to be much more subtle than
was expected, see e. g. [18,64].
taken in the sense of Whitney [29,119,136,143,144], also

see [20,24,31,55].
It follows that the occurrence of normally hyperbolic
invariant tori carrying (Diophantine) quasi-periodic ow
is typical for families of systems with suciently many
parameters, where this occurrence has positive measure
in parameter space. In fact, if the number of parameters
equals the dimension of the tori, the geometry as sketched
in Fig. 4 carries over in a dieomorphic way.
Dissipative KAM Theory
Many remarks following Subsect. Classical KAM Theory and Theorem 3 also hold here.
In cases where the system is degenerate, for instance
because there is a lack of parameters, a path formalism
can be invoked, where the parameter path is required to
n , see
be a generic subfamily of the Diophantine set R;
Fig. 4. This amounts to the Rssmann non-degeneracy,
that still gives positive measure of quasi-periodicity in
the parameter space, compare with [24,31] and references therein.
In the dissipative case the KAM theorem gives rise to
families of quasi-periodic attractors in a typical way.
This is of importance in center manifold reductions of
innite dimensional dynamics as, e. g., in uid mechanics [125,126]. In Sect. Transition to Chaos and Turbulence we shall return to this subject.
As already noted by Moser [105,106], KAM theory extends

outside the world of Hamiltonian systems, like to volume
preserving systems, or to equivariant or reversible systems.
This also holds for the class of general smooth systems,
often called dissipative. In fact, the KAM theorem allows
for a Lie algebra proof, that can be used to cover all these
special cases [24,29,31,45]. It turns out that in many cases
parameters are needed for persistent occurrence of (Diophantine) quasi-periodic tori.
As an example we now consider the dissipative setting,
where we discuss a parametrized system with normally hyperbolic invariant n-tori carrying quasi-periodic motion.
From [73] it follows that this is a persistent situation and
that, up to a smooth (in this case of class Ck for large k)
dieomorphism, we can restrict to the case where T n is
the phase space. To x thoughts we consider the smooth
system
' D !() C " f ('; ; ")
D0;

(8)
where 2 Rn is a multi-parameter. The results of the

classical KAM theorem regarding (6)" largely carry over
to (8);" .
Now, for " D 0 the product of phase space and parameter space as an open subset of T n Rn is completely foliated by invariant n-tori and since the perturbation does
not concern the -equation,

this foliation is persistent.
The interest is with the dynamics on the resulting invariant
tori that remains parallel after the perturbation; compare
with the setting of Theorem 3. As just stated, KAM theory here gives a solution similar to the Hamiltonian case.
The analogue of the Kolmogorov non-degeneracy condition here is that the frequency map 7! !() is a (local)
dieomorphism. Then, in the spirit of Theorem 3, we state
that the system (8);" is smoothly conjugated to (8);0 ,
as before, provided that the map ! is co-restricted to the
n . Again the smoothness has to be
Diophantine set R;
Remark
Lower Dimensional Tori

We extend the above approach to the case of lower dimensional tori, i. e., where the dynamics transversal to the
tori is also taken into account. We largely follow the setup of [29,45] that follows Moser [106]. Also see [24,31]
and references therein. Changing notation a little, we now
consider the phase space T n Rm D fx(mod 2); yg,
as well a parameter space fg D P Rs . We consider
a C 1 -family of vector elds X(x; y; ) as before, having
T n f0g T n Rm as an invariant n-torus for D
0 2 P.
x D !() C f (y; )
y D () y C g(y; )
(9)
D0;

with f (y; 0 ) D O(jyj) and g(y; 0 ) D O(jyj2 ), so we assume the invariant torus to be of Floquet type.
The system X D X(x; y; ) is integrable in the sense
that it is T n -symmetric, i. e., x-independent [29]. The interest is with the fate of the invariant torus T n f0g and
4529
4530
its parallel dynamics under small perturbation to a system
X D X(x;
y; ) that no longer needs to be integrable.
Consider the smooth mappings ! : P ! Rn and
: P ! gl(m; R). To begin with we restrict to the case
where all eigenvalues of (0 ) are simple and nonzero. In
general for such a matrix 2 gl(m; R), let the eigenvalues be given by 1 i1 ; : : : ; N 1 i N 1 and 1 ; : : : ; N 2 ,
where all j ; j and j are real and hence m D 2N1 C N2 .
Also consider the map spec : gl(m; R) ! R2N 1 CN 2 , given
by 7! (; ; ). Next to the internal frequency vector
! 2 Rn , we also have the vector 2 R N 1 of normal frequencies.
The present analogue of Kolmogorov non-degeneracy is the BroerHuitemaTakens (BHT) non-degeneracy
condition [29,127], which requires that the product map
! (spec) : P ! Rn gl(m; R) at D 0 has a surjective derivative and hence is a local submersion [72].
Furthermore, we need Diophantine conditions on
both the internal and the normal frequencies, generalizing (7). Given > n 1 and > 0, it is required for all
k 2 Z n n f0g and all ` 2 Z N 1 with j`j 2 that
jhk; !i C h`; ij jkj :
(10)
Inside Rn R N 1 D f!; g this yields a Cantor set as before (compare Fig. 4). This set has to be pulled back along
the submersion ! (spec) , for examples see Subsects. (n 1)-Tori and Quasi-periodic Bifurcations
below.
The KAM theorem for this setting is quasi-periodic stability of the n-tori under consideration, as in Subsect. Dissipative KAM Theory, yielding typical examples where
quasi-periodicity has positive measure in parameter space.
In fact, we get a little more here, since the normal linear
behavior of the n-tori is preserved by the Whitney smooth
conjugations. This is expressed as normal linear stability,
which is of importance for quasi-periodic bifurcations, see
Subsect. Quasi-periodic Bifurcations below.
Remark
A more general set-up of the normal stability theory [45] adapts the above to the case of non-simple (multiple) eigenvalues. Here the BHT non-degeneracy condition is formulated in terms of versal unfolding of the matrix (0 ) [7]. For possible conditions under which vanishing eigenvalues are admissible
see [29,42,69] and references therein.
This general set-up allows for a structure preserving formulation as mentioned earlier, thereby including the Hamiltonian and volume preserving case, as
well as equivariant and reversible cases. This allows
us, for example, to deal with quasi-periodic versions

of the Hamiltonian and the reversible Hopf bifurcation [38,41,42,44].
The Parameterized KAM Theory discussed here a priori needs many parameters. In many cases the parameters are distinguished in the sense that they are
given by action variables, etc. For an example see Subsect. (n 1)-Tori on Hamiltonian (n 1)-tori Also
see [127] and [24,31] where the case of Rssmann nondegeneracy is included. This generalizes a remark at the
end of Subsect. Dissipative KAM Theory.
Global KAM Theory
We stay in the Hamiltonian setting, considering Lagrangian invariant n-tori as these occur in a Liouville integrable system with n degrees of freedom. The union of
these tori forms a smooth T n -bundle f : M ! B (where
we leave out all singular bers). It is known that this bundle can be non-trivial [56,60] as can be measured by monodromy and Chern class. In this case global action angle
variables are not dened. This non-triviality, among other
things, is of importance for semi-classical versions of the
classical system at hand, in particular for certain spectrum
defects [57,62,134,135], for more references also see [24].
Restricting to the classical case, the problem is what
happens to the (non-trivial) T n -bundle f under small,
non-integrable perturbation. From the classical KAM theory, see Subsect. Classical KAM Theory we already know
that on trivializing charts of f Diophantine quasi-periodic
n-tori persist. In fact, at this level, a Whitney smooth conjugation exists between the integrable system and its perturbation, which is even Gevrey regular [136]. It turns out
that these local KAM conjugations can be glued together
so to obtain a global conjugation at the level of quasi-periodic tori, thereby implying global quasi-periodic stability [43]. Here we need unicity of KAM tori, i. e., independence of the action-angle chart used in the classical KAM
theorem [26]. The proof uses the integer ane structure
on the quasi-periodic tori, which enables taking convex
combinations of the local conjugations subjected to a suitable partition of unity [72,129]. In this way the geometry
of the integrable bundle can be carried over to the nearlyintegrable one.
The classical example of a Liouville integrable system with non-trivial monodromy [56,60] is the spherical pendulum, which we now briey revisit. The
conguration space is S2 D fq 2 R3 j hq; qi D 1g and
the phase space T S2 f(q; p) 2 R6 j hq; qi D 1 and
hq; pi D 0g. The two integrals I D q1 p2 q2 p1 (angular momentum) and E D 12 hp; pi C q3 (energy) lead to
nearly integrable case, so where the axial symmetry is

broken. Also compare [24] and many of its references.
The global conjugations of [43] are Whitney smooth
(even Gevrey regular [136]) and near the identity map
in the C 1 -topology [72]. Geometrically speaking these
dieomorphisms also are T n -bundle isomorphisms
between the unperturbed and the perturbed bundle, the
basis of which is a Cantor set of positive measure.
Splitting of Separatrices

Figure 5
Range of the energy-momentum map of the spherical pendulum
an energy momentum
map EM : T S2 ! R 2 , given by

(q; p) 7! (I; E) D q1 p2 q2 p1 ; 12 hp; pi C q3 . In Fig. 5
we show the image of the map EM. The shaded area B
consists of regular values, the ber above which is a Lagrangian two-torus; the union of these gives rise to a bundle f : M ! B as described before, where f D EMj M .
The motion in the two-tori is a superposition of Huygens rotations and pendulum-like swinging, and the nonexistence of global action angle variables reects that the
three interpretations of rotating oscillation, oscillating
rotation and rotating rotation cannot be reconciled in
a consistent way. The singularities of the bration include
the equilibria (q; p) D ((0; 0; 1); (0; 0; 0)) 7! (I; E) D
(0; 1). The boundary of this image also consists of singular points, where the ber is a circle that corresponds
to Huygens horizontal rotations of the pendulum. The
ber above the upper equilibrium point (I; E) D (0; 1) is
a pinched torus [56], leading to non-trivial monodromy,
in a suitable bases of the period lattices, given by

1
0
1
1

2 GL(2; R) :
The question here is what remains of the bundle f when

the system is perturbed. Here we observe that locally Kolmogorov non-degeneracy is implied by the non-trivial
monodromy [114,122]. From [43,122] it follows that the
non-trivial monodromy can be extended in the perturbed
case.
Remark
The case where this perturbation remains integrable is
covered in [95], but presently the interest is with the
KAM theory does not predict the fate of close-to-resonant

tori under perturbations. For fully resonant tori the phenomenon of frequency locking leads to the destruction
of the torus under (suciently rich) perturbations, and
other resonant tori disintegrate as well. In the case of a single resonance between otherwise Diophantine frequencies
the perturbation leads to quasi-periodic bifurcations, cf.
Sect. Transition to Chaos and Turbulence.
While KAM theory concerns the fate of most trajectories and for all times, a complementary theorem has been
obtained in [93,109,110,113]. It concerns all trajectories
and states that they stay close to the unperturbed tori for
long times that are exponential in the inverse of the perturbation strength. For trajectories starting close to surviving tori the diusion is even superexponentially slow,
cf. [102,103]. Here a form of smoothness exceeding the
mere existence of innitely many derivatives of the Hamiltonian is a necessary ingredient, for nitely dierentiable
Hamiltonians one only obtains polynomial times.
Solenoids, which cannot be present in integrable systems, are constructed for generic Hamiltonian systems
in [16,94,98], yielding the simultaneous existence of representatives of all homeomorphy-classes of solenoids. Hyperbolic tori form the core of a construction proposed
in [5] of trajectories that venture o to distant points of
the phase space. In the unperturbed system the union
of a family of hyperbolic tori, parametrized by the actions conjugate to the toral angles, form a normally hyperbolic manifold. The latter is persistent under perturbations, cf. [73,100], and carries a Hamiltonian ow with
fewer degrees of freedom. The main dierence between integrable and non-integrable systems already occurs for periodic orbits.
Periodic Orbits
A sharp dierence to dissipative systems is that it is generic
for hyperbolic periodic orbits on compact energy shells in
Hamiltonian systems to have homoclinic orbits, cf. [1] and
references therein. For integrable systems these form to-
4531
4532
gether a pinched torus, but under generic perturbations

the stable and unstable manifold of a hyperbolic periodic
orbit intersect transversely. It is a nontrivial task to actually check this genericity condition for a given non-integrable perturbation, a rst-order condition going back
to Poincar requires the computation of the so-called
Melnikov integral, see [66,137] for more details. In two
degrees of freedom normalization leads to approximations that are integrable to all orders, which implies that
the Melnikov integral is a at function. In the real analytic case the Melnikov criterion is still decisive in many
examples [65].
Genericity conditions are traditionally formulated in
the universe of smooth vector elds, and this makes the
whole class of analytic vector elds appear to be nongeneric. This is an overly pessimistic view as the conditions
dening a certain class of generic vector elds may certainly be satised by a given analytic system. In this respect
it is interesting that the generic properties may also be formulated in the universe of analytic vector elds, see [28]
for more details.
(n 1)-Tori
The (n 1)-parameter families of invariant (n 1)-tori
organize the dynamics of an integrable Hamiltonian system in n degrees of freedom, and under small perturbations the parameter space of persisting analytic tori is
Cantorized. This still allows for a global understanding of
a substantial part of the dynamics, but also leads to additional questions.
A hyperbolic invariant torus T n1 has its Floquet exponents o the imaginary axis. Note that T n1 is not
a normally hyperbolic manifold. Indeed, the normal linear
behavior involves the n 1 zero eigenvalues in the direction of the parametrizing actions as well; similar to (9) the
format
x D !(y) C O(y) C O(z2 )
y D O(y) C O(z3 )
z D (y)z C O(z2 )
in Floquet coordinates yields an x-independent matrix
that describes the symplectic normal linear behavior,
cf. [29]. The union fz D 0g over the family of (n 1)-tori
is a normally hyperbolic manifold and constitutes the center manifold of T n1 . Separatrices splitting yields the dividing surfaces in the sense of Wiggins et al. [138].
The persistence of elliptic tori under perturbation
from an integrable system involves not only the internal
frequencies of T n1 , but also the normal frequencies. Next
to the internal resonances the necessary Diophantine conditions (10) exclude the normal-internal resonances
hk; !i D j
(11)
hk; !i D 2 j
(12)
hk; !i D i C j
(13)
hk; !i D i j :
(14)
The rst three resonances lead to the quasi-periodic

center-saddle bifurcation studied in Sect. Transition to
Chaos and Turbulence, the frequency-halving (or quasiperiodic period doubling) bifurcation and the quasi-periodic Hamiltonian Hopf bifurcation, respectively. The resonance (14) generalizes an equilibrium in 1 : 1 resonance
whence T n1 persists and remains elliptic, cf. [78]. When
passing through resonances (12) and (13) the lower-dimensional tori lose ellipticity and acquire hyperbolic Floquet exponents. Elliptic (n 1)-tori have a single normal
frequency whence (11) and (12) are the only normal-internal resonances. See [35] for a thorough treatment of the
ensuing possibilities.
The restriction to a single normal-internal resonance
is dictated by our present possibilities. Indeed, already the
bifurcation of equilibria with a fourfold zero eigenvalue
leads to unfoldings that simultaneously contain all possible normal resonances. Thus, a satisfactory study of such
tori which already may form one-parameter families in integrable Hamiltonian systems with ve degrees of freedom
has to await further progress in local bifurcation theory.
Transition to Chaos and Turbulence
One of the main interests over the second half of the twentieth century has been the transition between orderly and
complicated forms of dynamics upon variation of either
initial states or of system parameters. By orderly we here
mean equilibrium and periodic dynamics and by complicated quasi-periodic and chaotic dynamics, although we
note that only chaotic dynamics is associated to unpredictability, e. g. see [27]. As already discussed in the introduction systems like a forced nonlinear oscillator or the
planar three-body problem exhibit coexistence of periodic,
quasi-periodic and chaotic dynamics, also compare with
Fig. 1.
Similar remarks go for the onset of turbulence in uid
dynamics. Around 1950 this led to the scenario of Hopf
LandauLifschitz [75,76,83,84], which roughly amounts
to the following. Stationary uid motion corresponds to
an equilibrium point in an 1-dimensional state space
of velocity elds. The rst transition is a Hopf bifurcation [66,75,82], where a periodic solution branches o.
In a second transition of similar nature a quasi-periodic
two-torus branches o, then a quasi-periodic three-torus,
etc. The idea is that the motion picks up more and more
frequencies and thus obtains an increasingly complicated
power spectrum. In the early 1970s this idea was modied
in the RuelleTakens route to turbulence, based on the observation that, for ows, a three-torus can carry chaotic
(or strange) attractors [112,126], giving rise to a broad
band power spectrum. By the quasi-periodic bifurcation
theory [24,29,31] as sketched below these two approaches
are unied in a generic way, keeping track of measure theoretic aspects. For general background in dynamical systems theory we refer to [27,79].
Another transition to chaos was detected in the
quadratic family of interval maps
f (x) D x(1 x) ;
see [58,99,101], also for a holomorphic version. This transition consists of an innite sequence of period doubling
bifurcations ending up in chaos; it has several universal
aspects and occurs persistently in families of dynamical
systems. In many of these cases also homoclinic bifurcations show up, where sometimes the transition to chaos
is immediate when parameters cross a certain boundary,
for general theory see [13,14,30,117]. There exist quite
a number of case studies where all three of the above scenarios play a role, e. g., see [32,33,46] and many of their
references.
Quasi-periodic Bifurcations
For the classical bifurcations of equilibria and periodic
orbits, the bifurcation sets and diagrams are generally
determined by a classical geometry in the product of
phase space and parameter space as already established
by, e. g., [8,133], often using singularity theory. Quasi-periodic bifurcation theory concerns the extension of these
bifurcations to invariant tori in nearly-integrable systems,
e. g., when the tori lose their normal hyperbolicity or when
certain (strong) resonances occur. In that case the dense
set of resonances, also responsible for the small divisors,
leads to a Cantorization of the classical geometries obtained from Singularity Theory [29,35,37,38,39,41,44,45,
48,49,67,68,69], also see [24,31,52,55]. Broadly speaking,
one could say that in these cases the Preparation Theorem [133] is partly replaced by KAM theory. Since the KAM
theory has been developed in several settings with or without preservation of structure, see Sect. KAM Theory: An
Overview, for the ensuing quasi-periodic bifurcation theory the same holds.
Hamiltonian Cases To x thoughts we start with an example in the Hamiltonian setting, where a robust model
for the quasi-periodic center-saddle bifurcation is given by
H!1 ;!2 ;;" (I; '; p; q)
D !1 I 1 C !2 I 2 C
1 2
p C V (q) C " f (I; '; p; q)
2
(15)
with V (q) D 13 q3 q, compare with [67,69]. The unperturbed (or integrable) case " D 0, by factoring out the
T 2 -symmetry, boils down to a standard center-saddle bifurcation, involving the fold catastrophe [133] in the potential function V D V (q). This results in the existence
of two invariant two-tori, one elliptic and the other hyperbolic. For 0 j"j
1 the dense set of resonances complicates this scenario, as sketched in Fig. 6, determined by the
Diophantine conditions
jhk; !ij jkj ;
jhk; !i C `(q)j jkj

for q < 0 ;
(16)
for q > 0
for all kp
2 Z n n f0g and for all ` 2 Z with j`j 2. Here
(q) D 2q is the normal frequency of the elliptic torus
p
given by q D for > 0. As before, (cf. Sects. Invariant Curves of Planar Dieomorphisms, KAM Theory:
An Overview), this gives a Cantor set of positive measure [24,29,31,45,69,105,106].
For 0 < j"j
1 Fig. 6 will be distorted by a nearidentity dieomorphism; compare with the formulations
of the Theorems 3 and 4. On the Diophantine Cantor set
the dynamics is quasi-periodic, while in the gaps generically there is coexistence of periodicity and chaos, roughly
comparable with Fig. 1, at left. The gaps at the border furthermore lead to the phenomenon of parabolic resonance,
cf. [86].
Similar programs exist for all cuspoid and umbilic
catastrophes [37,39,68] as well as for the Hamiltonian
Hopf bifurcation [38,44]. For applications of this approach
see [35]. For a reversible analogue see [41]. As so often
within the gaps generically there is an innite regress of
smaller gaps [11,35]. For theoretical background we refer
to [29,45,106], for more references also see [24].
Dissipative Cases In the general dissipative case we basically follow the same strategy. Given the standard bifurcations of equilibria and periodic orbits, we get more complex situations when invariant tori are involved as well.
4533
4534

Figure 6
Sketch of the Cantorized Fold, as the bifurcation set of the quasiperiodic center-saddle bifurcation for n D 2 [67], where the horizontal axis indicates the frequency ratio !2 : !1 , cf. (15). The
lower part of the figure corresponds to hyperbolic tori and the
upper part to elliptic ones. See the text for further interpretations
The simplest examples are the quasi-periodic saddle-node

and quasi-periodic period doubling [29] also see [24,31].
To illustrate the whole approach let us start from
the Hopf bifurcation of an equilibrium point of a vector eld [66,75,82,116] where a hyperbolic point attractor loses stability and branches o a periodic solution,
cf. Subsect. Dissipative Perturbations. A topological normal form is given by

y1
y2

D

y1
y2

y12
y22
y1
y2
(17)
where y D (y1 ; y2 ) 2 R2 , ranging near (0; 0). In this representation usually one xes D 1 and lets D (near
0) serve as a (bifurcation) parameter, classifying modulo
topological equivalence. In polar coordinates (17) so gets
the form
' D 1 ;
r D r r3 :
Figure 7 shows an amplitude response diagram (often
called the bifurcation diagram). Observe the occurrence
of the attracting periodic solution for > 0 of amplitude
p
.
Let us briey consider the Hopf bifurcation for xed
points of dieomorphisms. A simple example has the form
P(y) D e2 (Ci ) y C O(jyj2 ) ;
(18)
y 2 C R2 , near 0. To start with is considered a constant, such that is not rational with denominator less

Figure 7
Bifurcation diagram of the Hopf bifurcation
than ve, see [7,132], and where O(jyj2 ) should contain

generic third order terms. As before, we let D serve
as a bifurcation parameter, varying near 0. On one side of
the bifurcation value D 0, this system has by normal hyperbolicity and [73], an invariant circle. Here, due to the
invariance of the rotation numbers of the invariant circles,
no topological stability can be obtained [111]. Still this bifurcation can be characterized by many persistent properties. Indeed, in a generic two-parameter family (18), say
with both and as parameters, the periodicity in the parameter plane is organized in resonance tongues [7,34,82].
(The tongue structure is hardly visible when only one parameter, like , is used.) If the dieomorphism is the return map of a periodic orbit for ows, this bifurcation produces an invariant two-torus. Usually this counterpart for
ows is called NemarkSacker bifurcation. The periodicity as it occurs in the resonance tongues, for the vector
eld is related to phase lock. The tongues are contained in
gaps of a Cantor set of quasi-periodic tori with Diophantine frequencies. Compare the discussion in Subsect. Circle Maps, in particular also regarding the Arnold family
and Fig. 3. Also see Sect. KAM Theory: An Overview and
again compare with [115].
Quasi-periodic versions exist for the saddle-node, the
period doubling and the Hopf bifurcation. Returning to
the setting with T n Rm as the phase space, we remark
that the quasi-periodic saddle-node and period doubling
already occur for m D 1, or in an analogous center manifold. The quasi-periodic Hopf bifurcation needs m 2.
We shall illustrate our results on the latter of these cases,
compare with [19,31]. For earlier results in this direction
see [52]. Our phase space is T n R2 D fx(mod 2); yg,
where we are dealing with the parallel invariant torus
T n f0g. In the integrable case, by T n -symmetry we can
reduce to R2 D fyg and consider the bifurcations of relative equilibria. The present interest is with small non-integrable perturbations of such integrable models.
We now discuss the quasi-periodic Hopf bifurcation [17,29], largely following [55]. The unperturbed, integrable family X D X (x; y) on T n R2 has the form
X (x; y)
D [!() C f (y; )]@x C [()y C g(y; )]@ y ;
(19)
were f D O(jyj) and g D O(jyj2 ) as before. Moreover
2 P is a multi-parameter and ! : P ! Rn and : P !
gl(2; R) are smooth maps. Here we take

() ()
() D
;
()
()
which makes the @ y component of (19) compatible with
the planar Hopf family (17). The present form of Kolmogorov non-degeneracy is BroerHuitemaTakens stability [29,42,45], requiring that there is a subset P on
which the map
2 P 7! (!(); ()) 2 Rn gl(2; R)
is a submersion. For simplicity we even assume that is
replaced by
(!; (; )) 2 Rn R2 :
Observe that if the non-linearity g satises the well-known
Hopf non-degeneracy conditions, e. g., compare [66,82],
then the relative equilibrium y D 0 undergoes a standard
planar Hopf bifurcation as described before. Here again
plays the role of bifurcation parameter and a closed orbit branches o at D 0. To x thoughts we assume that
y D 0 is attracting for < 0. and that the closed orbit occurs for > 0, and is attracting as well. For the integrable
family X, qualitatively we have to multiply this planar scenario with T n , by which all equilibria turn into invariant
attracting or repelling n-tori and the periodic attractor into
an attracting invariant (n C 1)-torus. Presently the question is what happens to both the n- and the (n C 1)-tori,
when we apply a small near-integrable perturbation.
The story runs much like before. Apart from the BHT
non-degeneracy condition we require Diophantine conditions (10), dening the Cantor set
(2)
D f(!; (; )) 2 j jhk; !i C `j jkj ;
;
8k 2 Z n n f0g ; 8` 2 Z with j`j 2g ;
(20)
(2)
Rn R2
In Fig. 8 we sketch the intersection of ;

Figure 8
(2)
Planar section of the Cantor set ;
with a plane f!g R2 for a Diophantine (internal) frequency vector !, cf. (7).
From [17,29] it now follows that for any family X
on T n R2 P, suciently near X in the C 1 -topology
a near-identity C 1 -dieomorphism : T n R2 !
T n R2 exists, dened near T n f0g , that conju(2)
gates X to X when further restricting to T n f0g ;
.
So this means that the Diophantine quasi-periodic invariant n-tori are persistent on a dieomorphic image of the
(2)
, compare with the formulations of the
Cantor set ;
Theorems 3 and 4.
Similarly we can nd invariant (n C 1)-tori. We rst
have to develop a T nC1 symmetric normal form approximation [17,29] and Normal Forms in Perturbation Theory. For this purpose we extend the Diophantine conditions (20) by requiring that the inequality holds for all
j`j N for N D 7. We thus nd another large Cantor set,
again see Fig. 8, where Diophantine quasi-periodic invariant (n C 1)-tori are persistent. Here we have to restrict to
> 0 for our choice of the sign of the normal form coecient, compare with Fig. 7.
In both the cases of n-tori and of (n C 1)-tori, the
nowhere dense subset of the parameter space containing
the tori can be fattened by normal hyperbolicity to open
subsets. Indeed, the quasi-periodic n- and (n C 1)-tori are
innitely normally hyperbolic [73]. Exploiting the normal
form theory [17,29] and Normal Forms in Perturbation
Theory to the utmost and using a more or less standard
contraction argument [17,53], a fattening of the parameter
domain with invariant tori can be obtained that leaves out
only small bubbles around the resonances, as sketched
4535
4536
As an example consider a family of maps that undergoes a generic quasi-periodic Hopf bifurcation from circle to two-torus. It turns out that here the Cantorized fold
of Fig. 6 is relevant, where now the vertical coordinate is
a bifurcation parameter. Moreover compare with Fig. 3,
where also variation of " is taken into account. The Cantor
set contains the quasi-periodic dynamics, while in the gaps
we can have chaos, e. g., in the form of Hnon like strange
attractors [46,112]. A fattening process as explained above,
also can be carried out here.
Future Directions

Figure 9
Fattening by normal hyperbolicity of a nowhere dense parameter set with invariant n-tori in the perturbed system. The curve
H is the Whitney smooth (even Gevrey regular [136]) image
of the -axis in Fig. 8. H interpolates the Cantor set Hc
that contains the non-hyperbolic Diophantine quasi-periodic in(2)
, see (20). To the points
variant n-tori, corresponding to ;
1;2 2 Hc discs A 1;2 are attached where we find attracting
normally hyperbolic n-tori and similarly in the discs R 1;2 repelling ones. The contact between the disc boundaries and H
is infinitely flat [17,29]
and explained in Fig. 9 for the n-tori. For earlier results in

the same spirit in a case study of the quasi-periodic saddlenode bifurcation see [49,50,51], also compare with [11].
A Scenario for the Onset of Turbulence
Generally speaking, in many settings quasi-periodicity
constitutes the order in between chaos [31]. In the Hopf
LandauLifschitzRuelleTakens scenario [76,83,84,126]
we may consider a sequence of typical transitions as
given by quasi-periodic Hopf bifurcations, starting with
the standard Hopf or HopfNemarkSacker bifurcation
as described before. In the gaps of the Diophantine Cantor sets generically there will be coexistence of periodicity, quasi-periodicity and chaos in innite regress. As said
earlier, period doubling sequences and homoclinic bifurcations may accompany this.
One important general issue is the mathematical characterization of chaos and ergodicity in dynamical systems,
in conservative, dissipative and in other settings. This is
a tough problem as can already be seen when considering
two-dimensional dieomorphisms. In particular we refer
to the still unproven ergodicity conjecture of [9] and to the
conjectures around Hnon like attractors and the principle Hnon everywhere, compare with [22,32]. For a discussion see Subsect. A Scenario for the Onset of Turbulence. In higher dimension this problem is even harder to
handle, e. g., compare with [46,47] and references therein.
In the conservative case a related problem concerns a better understanding of Arnold diusion.
Somewhat related to this is the analysis of dynamical
systems without an explicit perturbation setting. Here numerical and symbolic tools are expected to become useful
to develop computer assisted proofs in extended perturbation settings, diagrams of Lyapunov exponents, symbolic
dynamics, etc. Compare with [128]. Also see [46,47] for
applications and further reference. This part of the theory
is important for understanding concrete models, that often are not given in perturbation format.
Regarding nearly-integrable Hamiltonian systems,
several problems have to be considered. Continuing the
above line of thought, one interest is the development of
Hamiltonian bifurcation theory without integrable normal
form and, likewise, of KAM theory without action angle coordinates [87]. One big related issue also is to develop KAM
theory outside the perturbation format.
The previous section addressed persistence of Diophantine tori involved in a bifurcation. Similar to Cremers example in Subsect. Cremers Example in Hermans Version the dynamics in the gaps between persistent tori displays new phenomena. A rst step has been
made in [86] where internally resonant parabolic tori involved in a quasi-periodic Hamiltonian pitchfork bifurcation are considered. The resulting large dynamical instabilities may be further amplied for tangent (or at)
parabolic resonances, which fail to satisfy the iso-energetic

non-degeneracy condition.
The construction of solenoids in [16,94] uses elliptic periodic orbits as starting points, the simplest example being the result of a period-doubling sequence. This
construction should carry over to elliptic tori, where normal-internal resonances lead to encircling tori of the same
dimension, while internal resonances lead to elliptic tori
of smaller dimension and excitation of normal modes increases the torus dimension. In this way one might be able
to construct solenoid-type invariant sets that are limits of
tori with varying dimension.
Concerning the global theory of nearly-integrable
torus bundles [43], it is of interest to understand the effects of quasi-periodic bifurcations on the geometry and
its invariants. Also it is of interest to extend the results
of [134] when passing to semi-classical approximations. In
that case two small parameters play a role, namely Plancks
constant as well as the distance away from integrability.
13.
14.
15.
16.
17.
18.
19.
20.
Bibliography
1. Abraham R, Marsden JE (1978) Foundations of Mechanics,
2nd edn. Benjamin
2. Arnold VI (1961) Small divisors I: On mappings of the circle
onto itself. Izv Akad Nauk SSSR Ser Mat 25:2186 (in Russian); English translation: Am Math Soc Transl Ser 2(46):213
284 (1965); Erratum: Izv Akad Nauk SSSR Ser Mat 28:479480
(1964, in Russian)
3. Arnold VI (1962) On the classical perturbation theory and the
stability problem of the planetary system. Dokl Akad Nauk
SSSR 145:487490
4. Arnold VI (1963) Proof of a theorem by A.N. Kolmogorov
on the persistence of conditionally periodic motions under
a small change of the Hamilton function. Russ Math Surv
18(5):936 (English; Russian original)
5. Arnold VI (1964) Instability of dynamical systems with several
degrees of freedom. Sov Math Dokl 5:581585
6. Arnold VI (1978) Mathematical Methods of Classical Mechanics, GTM 60. Springer, New York
7. Arnold VI (1983) Geometrical Methods in the Theory of Ordinary Differential Equations. Springer
8. Arnold VI (ed) (1994) Dynamical Systems V: Bifurcation Theory and Catastrophe Theory. Encyclopedia of Mathematical
Sciences, vol 5. Springer
9. Arnold VI, Avez A (1967) Problmes Ergodiques de la Mcanique classique, Gauthier-Villars; English edition: Arnold
VI, Avez A (1968) Ergodic problems of classical mechanics.
Benjamin
10. Arnold VI, Kozlov VV, Neishtadt AI (1988) Mathematical Aspects of Classical and Celestial Mechanics. In: Arnold VI (ed)
Dynamical Systems, vol III. Springer
11. Baesens C, Guckenheimer J, Kim S, MacKay RS (1991) Three
coupled oscillators: Mode-locking, global bifurcation and
toroidal chaos. Phys D 49(3):387475
12. Barrow-Green J (1997) Poincar and the Three Body Problem.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
In: History of Mathematics, vol 11. Am Math Soc, Providence;

London Math Soc, London
Benedicks M, Carleson L (1985) On iterations of 1 ax2 on
(1; 1). Ann Math 122:125
Benedicks M, Carleson L (1991) The dynamics of the Hnon
map. Ann Math 133:73169
Benettin G (2005) Physical applications of Nekhoroshev theorem and exponential estimates. In: Giorgilli A (ed) Hamiltonian dynamics theory and applications, Cetraro 1999, Lecture
Notes in Mathematics, vol 1861. Springer, pp 176
Birkhoff BD (1935) Nouvelles recherches sur les systemes dynamiques. Mem Pont Acad Sci Novi Lyncaei 1(3):85216
Braaksma BLJ, Broer HW (1987) On a quasi-periodic Hopf bifurcation. Ann Inst Henri Poincar, Anal non linaire 4(2):
115168
Bricmont J (1996) Science of chaos or chaos in science? In:
Gross PR, Levitt N, Lewis MW (eds) The Flight from Science
and Reason (New York, 1995), Ann New York Academy of Sciences, vol 775. New York Academy of Sciences, New York,
pp 131175; Also appeared in: Phys Mag 17:159208 (1995)
Broer HW (2003) Coupled Hopf-bifurcations: Persistent examples of n-quasiperiodicity determined by families of 3-jets.
Astrisque 286:223229
Broer HW (2004) KAM theory: the legacy of Kolmogorovs
1954 paper. Bull Am Math Soc (New Series) 41(4):507521
Broer HW, Huitema GB (1991) A proof of the isoenergetic
KAM -theorem from the ordinary one. J Differ Equ 90:5260
Broer HW, Krauskopf B (2000) Chaos in periodically driven systems. In Krauskopf B, Lenstra D (eds) Fundamental Issues of
Nonlinear Laser Dynamics. American Institute of Physics Conference Proceedings 548:3153
Broer HW, Roussarie R (2001) Exponential confinement of
chaos in the bifurcation set of real analytic diffeomorphisms.
In: Broer HW, Krauskopf B, Vegter G (eds) Global Analysis of
Dynamical Systems, Festschrift dedicated to Floris Takens for
his 60th birthday. Bristol and Philadelphia IOP, pp 167210
Broer HW, Sevryuk MB (2007) KAM Theory: quasi-periodicity
in dynamical systems. In: Broer HW, Hasselblatt B, Takens F
(eds) Handbook of Dynamical Systems, vol 3. North-Holland
(to appear)
Broer HW, Takens F (1989) Formally symmetric normal forms
and genericity. Dyn Rep 2:3660
Broer HW, Takens F (2007) Unicity of KAM tori. Ergod Theory
Dyn Syst 27:713724
Broer HW, Takens F (2008) Dynamical Systems and Chaos. To
be published by Epsilon Uitgaven
Broer HW, Tangerman FM (1986) From a differentiable to
a real analytic perturbation theory, applications to the Kupka
Smale theorems. Ergod Theory Dyn Syst 6:345362
Broer HW, Huitema GB, Takens F, Braaksma BLJ (1990) Unfoldings and bifurcations of quasi-periodic tori. In: Memoir AMS,
vol 421. Amer Math Soc, Providence
Broer HW, Dumortier F, van Strien SJ, Takens F (1991) Structures in dynamics, finite dimensional deterministic studies. In:
de Jager EM, van Groesen EWC (eds) Studies in Mathematical
Physics, vol II. North-Holland
Broer HW, Huitema GB, Sevryuk MB (1996) Quasi-Periodic Motions in Families of Dynamical Systems: Order amidst Chaos.
In: Lecture Notes in Mathematics, vol 1645. Springer
Broer HW, Sim C, Tatjer JC (1998) Towards global models
4537
4538
H
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
near homoclinic tangencies of dissipative diffeomorphisms.

Nonlinearity 11(3):667770
Broer HW, Sim C, Vitolo R (2002) Bifurcations and strange attractors in the Lorenz-84 climate model with seasonal forcing.
Nonlinearity 15(4):12051267
Broer HW, Golubitsky M, Vegter G (2003) The geometry of resonance tongues: a singularity theory approach. Nonlinearity
16:15111538
Broer HW, Hanmann H, Jorba , Villanueva J, Wagener
FOO (2003) Normal-internal resonances in quasi-periodically forced oscillators: a conservative approach. Nonlinearity
16:17511791
Broer HW, Hoveijn I, Lunter G, Vegter G (2003) Bifurcations
in Hamiltonian systems: Computing Singularities by Grbner
Bases. In: Lecture Notes in Mathematics, vol 1806. Springer
Broer HW, Hanmann H, You J (2005) Bifurcations of normally
parabolic tori in Hamiltonian systems. Nonlinearity 18:1735
1769
Broer HW, Hanmann H, Hoo J, Naudot V (2006) Nearly-integrable perturbations of the Lagrange top: applications of KAM
theory. In: Denteneer D, den Hollander F, Verbitskiy E (eds)
Dynamics & Stochastics: Festschrift in Honor of MS Keane Lecture Notes, vol 48. Inst. of Math. Statistics, pp 286303
Broer HW, Hanmann H, You J (2006) Umbilical torus bifurcations in Hamiltonian systems. J Differ Equ 222:233262
Broer HW, Naudot V, Roussarie R (2006) Catastrophe theory in
Dulac unfoldings. Ergod Theory Dyn Syst 26:135
Broer HW, Ciocci MC, Hanmann H (2007) The quasi-periodic
reversible Hopf bifurcation. In: Doedel E, Krauskopf B, Sanders
J (eds) Recent Advances in Nonlinear Dynamics: Theme section dedicated to Andr Vanderbauwhede. Intern J Bifurc
Chaos 17:26052623
Broer HW, Ciocci MC, Hanmann H, Vanderbauwhede A
(2009) Quasi-periodic stability of normally resonant tori. Phys
D 238:309318
Broer HW, Cushman RH, Fass F, Takens F (2007) Geometry
of KAM tori for nearly integrable Hamiltonian systems. Ergod
Theory Dyn Syst 27(3):725741
Broer HW, Hanmann H, Hoo J (2007) The quasi-periodic
Hamiltonian Hopf bifurcation. Nonlinearity 20:417460
Broer HW, Hoo J, Naudot V (2007) Normal linear stability of
quasi-periodic tori. J Differ Equ 232:355418
Broer HW, Sim C, Vitolo R (2008) The HopfSaddle-Node bifurcation for fixed points of 3D-diffeomorphisms, the Arnold
resonance web. Bull Belg Math Soc Simon Stevin 15:769787
Broer HW, Sim C, Vitolo R (2008) The HopfSaddle-Node bifurcation for fixed points of 3D-diffeomorphisms, analysis of
a resonance bubble. Phys D Nonlinear Phenom (to appear)
Broer HW, Hanmann H, You J (in preparation) On the destruction of resonant Lagrangean tori in Hamiltonian systems
Chenciner A (1985) Bifurcations de points fixes elliptiques I,
courbes invariantes. Publ Math IHS 61:67127
Chenciner A (1985) Bifurcations de points fixes elliptiques II,
orbites priodiques et ensembles de Cantor invariants. Invent
Math 80:81106
Chenciner A (1988) Bifurcations de points fixes elliptiques III,
orbites priodiques de petites priodes et limination rsonnante des couples de courbes invariantes. Publ Math IHS
66:591
Chenciner A, Iooss G (1979) Bifurcations de tores invariants.
Arch Ration Mech Anal 69(2):109198; 71(4):301306
53. Chow S-N, Hale JK (1982) Methods of Bifurcation Theory.

Springer
54. Chow S-N, Li C, Wang D (1994) Normal Forms and Bifurcation of Planar Vector Fields. Cambridge University Press,
Cambridge
55. Ciocci MC, Litvak-Hinenzon A, Broer HW (2005) Survey on dissipative KAM theory including quasi-periodic bifurcation theory based on lectures by Henk Broer. In: Montaldi J, Ratiu T
(eds) Geometric Mechanics and Symmetry: the Peyresq Lectures, LMS Lecture Notes Series, vol 306. Cambridge University Press, Cambridge, pp 303355
56. Cushman RH, Bates LM (1997) Global Aspects of Classical Integrable Systems. Birkhuser, Basel
57. Cushman RH, Dullin HR, Giacobbe A, Holm DD, Joyeux M,
Lynch P, Sadovski DA and Zhilinski BI (2004) CO2 molecule as
a quantum realization of the 1 : 1 : 2 resonant swing-spring
with monodromy. Phys Rev Lett 93:024302
58. Devaney RL (1989) An Introduction to Chaotic Dynamical Systems, 2nd edn. Addison-Wesley, Redwood City
59. Diacu F, Holmes P (1996) Celestial Encounters. The Origins of
Chaos and Stability. Princeton University Press, Princeton
60. Duistermaat JJ (1980) On global action-angle coordinates.
Commun Pure Appl Math 33:687706
61. Dumortier F, Roussarie R, Sotomayor J (1991) Generic 3-parameter families of vector fields, unfoldings of saddle, focus
and elliptic singularities with nilpotent linear parts. In: Dumortier F, Roussarie R. Sotomayor J, Zoladek H (eds) Bifurcations of Planar Vector Fields: Nilpotent Singularities and
Abelian Integrals. LNM 1480, pp 1164
62. Efstafhiou K (2005) Metamorphoses of Hamiltonian systems
with symmetries. LNM, vol 1864. Springer, Heidelberg
63. Fjoz J (2004) Dmonstration du thorme dArnold sur la
stabilit du systme plantaire (daprs Herman). Ergod Theory Dyn Syst 24:162
64. Gallavotti G, Bonetto F, Gentile G (2004) Aspects of Ergodic,
Qualitative and Statistical Theory of Motion. Springer
65. Gelfreich VG, Lazutkin VF (2001) Splitting of Separatrices: perturbation theory and exponential smallness. Russ Math Surv
56:499558
66. Guckenheimer J, Holmes P (1983) Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields. Springer
67. Hanmann H (1988) The quasi-periodic centre-saddle bifurcation. J Differ Equ 142:305370
68. Hanmann H (2004) Hamiltonian Torus Bifurcations Related
to Simple Singularities. In: Ladde GS, Medhin NG, Sambandham M (eds) Dynamic Systems and Applications, Atlanta
2003. Dynamic Publishers, pp 679685
69. Hanmann H (2007) Local and Semi-Local Bifurcations in
Hamiltonian Dynamical Systems Results and Examples. In:
Lecture Notes in Mathematics, vol 1893. Springer, Berlin
70. Herman M (1977) Mesure de Lebesgue et nombre de rotation. In: Palis J, do Carmo M (eds) Geometry and Topology. In:
Lecture Notes in Mathematics, vol 597. Springer, pp 271293
71. Herman MR (1979) Sur la conjugaison diffrentiable des diffomorphismes du cercle des rotations. Publ Math IHS
49:5233
72. Hirsch MW (1976) Differential Topology. Springer
73. Hirsch MW, Pugh CC, Shub M (1977) Invariant Manifolds. In:
Lecture Notes in Mathematics, vol 583. Springer
74. Hofer H, Zehnder E (1994) Symplectic invariants and Hamiltonian dynamics. Birkhuser
75. Hopf E (1942) Abzweigung einer periodischen Lsung von

einer stationren Lsung eines Differentialsystems. Ber MathPhys Kl Schs Akad Wiss Leipzig 94:122
76. Hopf E (1948) A mathematical example displaying features of
turbulence. Commun Appl Math 1:303322
77. Huygens C vres compltes de Christiaan Huygens, (1888
1950), vol 5, pp 241263 and vol 17, pp 156189. Martinus
Nijhoff, The Hague
78. de Jong HH (1999) Quasiperiodic breathers in systems of
weakly coupled pendulums: Applications of KAM theory
to classical and statistical mechanics. Ph D Thesis, Univ.
Groningen
79. Katok A, Hasselblatt B (1995) Introduction to the Modern
Theory of Dynamical Systems. Cambridge University Press,
Cambridge
80. Kolmogorov AN (1954) On the persistence of conditionally
periodic motions under a small change of the Hamilton function. Dokl Akad Nauk SSSR 98:527530 (in Russian); English
translation: Stochastic Behavior in Classical and Quantum
Hamiltonian Systems, Volta Memorial Conference (Como,
1977). In: Casati G, Ford J (eds) Lecture Notes in Physics,
vol 93. Springer, Berlin pp 5156 (1979); Reprinted in: Bai Lin
Hao (ed) Chaos. World Scientific, Singapore, pp 8186 (1984)
81. Kolmogorov AN (1957) The general theory of dynamical systems and classical mechanics. In: Gerretsen JCH, de Groot
J (eds) Proceedings of the International Congress of Mathematicians, vol 1 (Amsterdam, 1954), North-Holland, Amsterdam, pp 315333 (in Russian); Reprinted in: International
Mathematical Congress in Amsterdam, (1954) (Plenary Lectures). Fizmatgiz, Moscow, pp 187208 (1961); English translation as Appendix D in: Abraham RH (1967) Foundations of
Mechanics. Benjamin, New York, pp 263279; Reprinted as
Appendix in [1], pp 741757
82. Kuznetsov YA (2004) Elements of Applied Bifurcation Theory,
3rd edn. In: Applied Mathematical Sciences, vol 112. Springer,
New York
83. Landau LD (1944) On the problem of turbulence. Akad Nauk
44:339
84. Landau LD, Lifschitz EM (1959) Fluid Mechanics. Pergamon,
Oxford
85. Laskar J (1995) Large scale chaos and marginal stability in
the Solar System, XIth International Congress of Mathematical Physics (Paris, 1994). In: Iagolnitzer D (ed) Internat Press,
Cambridge, pp 75120
86. Litvak-Hinenzon A, Rom-Kedar V (2002) Parabolic resonances
in 3 degree of freedom near-integrable Hamiltonian systems.
Phys D 164:213250
87. de la Llave R, Gonzlez A, Jorba , Villanueva J (2005) KAM theory without action-angle variables. Nonlinearity 18:855895
88. Lochak P (1999) Arnold diffusion; a compendium of remarks
and questions. In: Sim C (ed) Hamiltonian systems with three
or more degrees of freedom (SAgar, 1995), NATO ASI Series C: Math Phys Sci, vol 533. Kluwer, Dordrecht, pp 168183
89. Lochak P, Marco J-P (2005) Diffusion times and stability exponents for nearly integrable analytic systems, Central Eur J
Math 3:342397
90. Lochak P, Neshtadt AI (1992) Estimates of stability time for
nearly integrable systems with a quasiconvex Hamiltonian.
Chaos 2:495499
91. Lukina O (2008) Geometry of torus bundles in Hamiltonian
systems, Ph D Thesis, Univ. Groningen
92. MacKay RS (1993) Renormalisation in area-preserving maps.

World Scientific
93. Marco J-P, Sauzin D (2003) Stability and instability for
Gevrey quasi-convex near-integrable Hamiltonian systems.
Publ Math Inst Hautes Etud Sci 96:199275
94. Markus L, Meyer KR (1980) Periodic orbits and solenoids
in generic Hamiltonian dynamical systems. Am J Math 102:
2592
95. Matveev VS (1996) Integrable Hamiltonian systems with
two degrees of freedom. Topological structure of saturated
neighborhoods of points of focus-focus and saddle-saddle
types. Sb Math 187:495524
96. McDuff D, Salamon D (1995) Introduction to Symplectic Geometry. Clarendon/Oxford University Press
97. Meyer KR, Hall GR (1992) Introduction to Hamiltonian Dynamical Systems and the N-Body Problem. In: Applied Mathematical Sciences, vol 90. Springer
98. Meiss JD (1986) Class renormalization: Islands around islands.
Phys Rev A 34:23752383
99. de Melo W, van Strien SJ (1991) One-Dimensional Dynamics.
Springer
100. Mielke A (1991) Hamiltonian and Lagrangian Flows on Center
Manifolds with Applications to Elliptic Variational Problems.
In: Lecture Notes in Mathematics, vol 1489. Springer
101. Milnor JW (2006) Dynamics in One Complex Variable, 3rd edn.
In: Ann. Math. Studies, vol 160. Princeton University Press,
Princeton
102. Morbidelli A, Giorgilli A (1995) Superexponential Stability of
KAM Tori. J Stat Phys 78:16071617
103. Morbidelli A, Giorgilli A (1995) On a connection between KAM
and Nekhoroshevs theorems. Physica D 86:514516
104. Moser JK (1962) On invariant curves of area-preserving mappings of an annulus. Nachr Akad Wiss Gttingen II, Math-Phys
Kl 1:120
105. Moser JK (1966) On the theory of quasiperiodic motions. SIAM
Rev 8(2):145172
106. Moser JK (1967) Convergent series expansions for quasi-periodic motions. Math Ann 169:136176
107. Moser JK (1968) Lectures on Hamiltonian systems. Mem Am
Math Soc 81:160
108. Moser JK (1973) Stable and random motions in dynamical systems, with special emphasis to celestial mechanics. In: Ann.
Math. Studies, vol 77. Princeton University Press, Princeton
109. Nekhoroshev NN (1977) An exponential estimate of the time
of stability of nearly-integrable Hamiltonian systems. Russ
Math Surv 32:165
110. Nekhoroshev NN (1985) An exponential estimate of the time
of stability of nearly integrable Hamiltonian systems II. In:
Oleinik OA (ed) Topics in Modern Mathematics, Petrovskii
Seminar No.5. Consultants Bureau, pp 158
111. Newhouse SE, Palis J, Takens F (1983) Bifurcations and stability of families of diffeomorphisms. Publ Math IHS 57:571
112. Newhouse SE, Ruelle D, Takens F (1978) Occurrence of
strange Axiom A attractors near quasi-periodic flows on T m ,
m 3. Commun Math Phys 64:3540
113. Niederman L (2004) Prevalence of exponential stability
among nearly-integrable Hamiltonian systems. Ergod Theory
Dyn Syst 24(2):593608
114. Nguyen Tien Zung (1996) Kolmogorov condition for integrable systems with focus-focus singularities. Phys Lett A
215(1/2):4044
4539
4540
H
115. Oxtoby J (1971) Measure and Category. Springer
116. Palis J, de Melo M (1982) Geometric Theory of Dynamical Systems. Springer
117. Palis J, Takens F (1993) Hyperbolicity & Sensitive Chaotic
Dynamics at Homoclinic Bifurcations. Cambridge University
Press, Cambridge
118. Poincar H (1980) Sur le problme des trois corps et les quations de la dynamique. Acta Math 13:1270
119. Pschel J (1982) Integrability of Hamiltonian systems on Cantor sets. Commun Pure Appl Math 35(5):653696
120. Pschel J (1993) Nekhoroshev estimates for quasi-convex
Hamiltonian systems. Math Z 213:187216
121. Pschel J (2001) A lecture on the classical KAM Theorem. In:
Proc Symp Pure Math 69:707732
122. Rink BW (2004) A Cantor set of tori with monodromy near
a focus-focus singularity. Nonlinearity 17:347356
123. Robinson C (1995) Dynamical Systems. CRC Press
124. Roussarie R (1997) Smoothness properties of bifurcation diagrams. Publ Mat 41:243268
125. Ruelle D (1989) Elements of Differentiable Dynamics and Bifurcation Theory. Academic Press
126. Ruelle D, Takens F (1971) On the nature of turbulence. Commun Math Phys 20:167192; 23:343344
127. Sevryuk MB (2007) Invariant tori in quasi-periodic non-autonomous dynamical systems via Hermans method. DCDS-A
18(2/3):569595
128. Sim C (2001) Global dynamics and fast indicators. In: Broer
HW, Krauskopf B, Vegter G (eds) Global Analysis of Dynamical Systems, Festschrift dedicated to Floris Takens for his 60th
birthday. IOP, Bristol and Philadelphia, pp 373390
129. Spivak M (1970) Differential Geometry, vol I. Publish or Perish
130. Takens F (1973) Introduction to Global Analysis. Comm. 2 of
the Math. Inst. Rijksuniversiteit Utrecht
131. Takens F (1974) Singularities of vector fields. Publ Math IHS
43:47100
132. Takens F (1974) Forced oscillations and bifurcations. In: Applications of Global Analysis I, Comm 3 of the Math Inst Rijksuniversiteit Utrecht (1974); In: Broer HW, Krauskopf B, Vegter G
(eds) Global Analysis of Dynamical Systems, Festschrift dedicated to Floris Takens for his 60th birthday. IOP, Bristol and
Philadelphia, pp 162
133. Thom R (1989) Structural Stability and Morphogenesis. An
Outline of a General Theory of Models, 2nd edn. AddisonWesley, Redwood City (English; French original)
134. Vu Ngo.c San (1999) Quantum monodromy in integrable systems. Commun Math Phys 203:465479
135. Waalkens H, Junge A, Dullin HR (2003) Quantum monodromy
in the two-centre problem. J Phys A Math Gen 36:L307-L314
136. Wagener FOO (2003) A note on Gevrey regular KAM theory
and the inverse approximation lemma. Dyn Syst 18:159163
137. Wiggins S (1990) Introduction to Applied Nonlinear Dynamical Systems and Chaos. Springer
138. Wiggins S, Wiesenfeld L, Jaffe C, Uzer T (2001) Impenetrable
barriers in phase-space. Phys Rev Lett 86(24):54785481
139. Yoccoz J-C (1983) C 1 -conjugaisons des diffomorphismes du
cercle. In: Palis J (ed) Geometric Dynamics, Proceedings, Rio
de Janeiro (1981) Lecture Notes in Mathematics, vol 1007,
pp 814827
140. Yoccoz J-C (1992) Travaux de Herman sur les tores invariants. In: Sminaire Bourbaki, vol 754, 19911992. Astrisque
206:311344
141. Yoccoz J-C (1995) Thorme de Siegel, nombres de Bruno et

polynmes quadratiques. Astrisque 231:388
142. Yoccoz J-C (2002) Analytic linearization of circle diffeomorphisms. In: Marmi S, Yoccoz J-C (eds) Dynamical Systems
and Small Divisors, Lecture Notes in Mathematics, vol 1784.
Springer, pp 125174
143. Zehnder E (1974) An implicit function theorem for small divisor problems. Bull Am Math Soc 80(1):174179
144. Zehnder E (1975) Generalized implicit function theorems
with applications to some small divisor problems, I and II.
Commun Pure Appl Math 28(1):91140; (1976) 29(1):49111

Hamiltonian Perturbation Theory (And Transition To Chaos) 4515

Uploaded by

Hamiltonian Perturbation Theory (And Transition To Chaos) 4515

Uploaded by

Hamiltonian Perturbation Theory (and Transition to Chaos)