Professional Documents
Culture Documents
Attilio Maccari
Author All books published by WILEY-VCH are carefully
produced. Nevertheless, authors, editors, and
Prof. Attilio Maccari publisher do not warrant the information
Via Alfredo Casella 3 contained in these books, including this book,
00013 Rome to be free of errors. Readers are advised to keep
Italy in mind that statements, data, illustrations,
procedural details or other items may
Cover Image: © Callista Images/Getty inadvertently be inaccurate.
Images
Library of Congress Card No.: applied for
Contents
12 Cosmology 187
12.1 Introduction 187
12.2 A New Field Equation 188
12.3 Exact Solution in the Robertson–Walker Metrics 191
12.4 Entropy Production 195
12.5 Conclusion 197
Conclusion 217
References 219
Index 235
xi
Foreword
Introduction
1.1 Introduction
Oscillations are a fundamental topic in physics. When a system is near its equilib-
rium point, it begins to oscillate, but if the displacement increases, then the nonlin-
ear terms are not negligible. The starting point is the differential equation for the
harmonic oscillator
d2 X
+ 𝜔2 X(t) = 0 (1.1)
dt
where X(t) is the displacement and ω the circular frequency. The most general solu-
tion is
where 𝜌 and 𝜃 are fixed by the initial conditions (the Cauchy problem)
and
( ̇ )
X0
tan𝜃 = (1.4)
𝜔X0
Now, we can consider a weakly nonlinear part in the differential Eq. (1.1) or, on the
contrary, a strongly nonlinear part but with small solutions. The first consequence is
that the amplitude and the phase are slowly varying with time, so we can introduce
another slow time
𝜏 = 𝜀q t (1.5)
where 𝜀 is a bookkeeping device and q is a rational number that will be chosen after-
wards. If we want to study the asymptotic solution behavior (t → ∞) and 𝜀 → 0, then
Asymptotic Perturbation Methods: For Nonlinear Differential Equations in Physics,
First Edition. Attilio Maccari.
© 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.
2 1 The Asymptotic Perturbation Method for Nonlinear Oscillators
𝜏 must assume finite values. So, we assume that an approximate solution is given by
X(t) = 2𝜌(𝜏) cos(−𝜔t + 𝜃(𝜏)) = (𝜌(𝜏) exp(−i𝜔t + i𝜃) + c.c.) (1.6)
or better
2𝜋 A 2
T= where 𝛺 = 𝜔 − 𝜌 (1.18)
𝛺 2𝜔
However, if
( )
10a2
b=− (1.19)
3𝜔2
the period does not change and is equal to the linear case period.
In this chapter, we want to extend this method and study a generalized Van der
Pol–Duffing oscillator in resonance with a periodic excitation
X(t) ̇
̈ + X(t) + f2 X 2 (t) + f3 X 3 (t) = 𝜀2 g0 X(t) ̇
+ g1 X(t)X(t) ̇
+ g2 X 2 (t)X(t) + 2𝜀2 f cos(t)
(1.27)
We demonstrate that, under appropriate conditions, a stable limit cycle appears
and calculate the relative approximate solution. Moreover, we derive sufficient
conditions for the existence of a doubly periodic motion when the fundamental
1.3 The Approximate Solution 5
The AP method we use to calculate the approximate solution was first developed in
Refs. [1, 2], and then in this section, we sketch the main steps of this perturbation
technique.
First of all, we now introduce a rational number
t = eq t (1.29)
where the rational number q will be fixed afterwards because it establishes to what
extent we can push the temporal asymptotic limit in such a way that the nonlinear
effects become consistent and not negligible. If t → ∞, then 𝜀 → 0, when 𝜏 assumes
a finite value.
If we take 𝜀 = 0 in Eqs. (1.26) and neglect nonlinear terms, we see that it admits
simple harmonic solutions X(t) = A exp(−it) + c. c., where A is a constant depend-
ing on initial conditions and c. c. stands for complex conjugate. Nonlinear effects
induce a modulation of the amplitude A and the appearance of higher harmonics.
The modulation is best described in terms of the rescaled variable t that accounts for
the need to look on larger time scales, to obtain a nonnegligible contribution from
the nonlinear term.
The assumed solution X(t) of (1.26) can be expressed by means of a power series
in the expansion parameter 𝜀, we formally write
∑
+∞
X(t) = 𝜀𝛾n 𝛹n (t, 𝜀) exp(−int) (1.30a)
n=−∞
with 𝛾 n = |n| for n ≠ 0, and 𝛾 0 = r is a positive number, which will be fixed later on;
in consequence of the reality of (1.30a)
The assumed solution (1.30a) can be considered a combination of the different har-
monics, solutions of the linear equation, i.e. of the equation obtained after neglecting
all the nonlinear terms, and the coefficients of this combination depend on 𝜏 and 𝜀.
Eq. (1.30a) can be written more explicitly
The functions 𝛹 n (t, 𝜀) depend on the parameter 𝜀, and we suppose that 𝛹 n ’s limit
for 𝜀 → 0 exists and is finite and, moreover, they can be expanded in power series of
𝜀, i.e.
∑
∞
𝛹n (𝜏; 𝜀) = 𝜀i 𝛹n(i) (𝜏) (1.31)
i=0
In the following, for simplicity, we use the abbreviations 𝛹n(0) = 𝛹n for n ≠ 1 and
𝛹1(0)
= 𝛹 for n = 1.
Note that the variable change (1.29) implies that
( )
(𝜕𝜓n exp(−int)) 𝜕𝜓
= −in𝜓n + 𝜀q n exp(−int) (1.32)
𝜕t 𝜕t
After inserting this expansion into Eq. (1.26), we obtain equations for every har-
monic and for a fixed order of approximation, which are right for the purpose of
determining the coefficients.
For n = 0, we obtain
𝜀r 𝛹0 + 2f2 |𝛹 |2 𝜀2 + o(𝜀4 , 𝜀r+2 ) = 0 (1.33a)
A correct balance of terms shows r = 2, and then we derive the following relation
𝜀2 𝛹0 = −2𝜀2 f2 (|𝛹 |2 ) + O(𝜀4 ) (1.33b)
For n = 2, taking into account Eq. (1.32), we have
−3𝜀2 𝛹2 + f2 𝛹 2 𝜀2 = −ig1 𝜀2 𝛹 2 + o(𝜀4 , 𝜀2+q ) (1.34a)
and then
f2 + ig1 2 2
𝜀2 𝛹 2 = 𝜀 𝛹 + o(𝜀4 ) (1.34b)
3
For n = 1, Eq. (1.26) yields for the right-hand side
d𝛹 1+q
−2i 𝜀 + 2f2 (𝛹0 𝛹 𝜀3 + 𝛹2 (c.c.(𝛹 ))𝜀3 ) + 3f3 |𝛹 |2 𝛹 𝜀3 (1.35a)
dt
and for the left-hand side
i𝜀3 g0 𝛹 − ig1 (𝛹0 𝛹 𝜀3 + 𝛹 𝛹2 (c.c.𝛹 )𝜀3 ) − i𝜀3 g2 |𝛹 |2 𝛹 + o(𝜀5 , 𝜀1+2q ) (1.35b)
If q = 2, the first term has the same magnitude order of nonlinear terms.
Taking into account Eqs. (1.33b) and (1.34b), we can derive a differential equation,
which involves only 𝛹 ,
d𝛹
= 𝛼1 𝛹 + (𝛽1 + i𝛽2 )|𝛹 |2 𝛹 (1.36)
dt
with
g0
𝛼1 = (1.37)
2
g2 g1 f2
𝛽1 = − (1.38)
2 2
g12 3 5
𝛽2 = − f3 + f22 (1.39)
6 2 3
1.3 The Approximate Solution 7
The term ddt𝛹2 in Eq. (1.50) can be eliminated taking into account that if we differ-
2
and Eq. (1.47c),(1.48), and (1.52a), then the final form of Eq. (1.46) is
d𝛹m
𝛼2 )𝛹m + (𝛽1 + 𝛽̃1 + i𝛽2 + i𝛽̃2 )|𝛹m |2 𝛹m + (̃
= (𝛼1 + ĩ g2 )|𝛹m |4 𝛹m
g1 + ĩ
dt
(1.53b)
with
( )
𝛼12g0 𝛼1
𝛼
̃2 = − + (1.54)
2 2
g1 g 1 3 5
𝛽̃1 = A1 + f2 B
̃1 + 1 B − g g2 − g f + g f 2 (1.55)
2 2 24 0 1 4 0 3 6 0 2
g1 g0 g2 3
𝛽̃2 = B ̃ −f B −f A − − g0 g1 f2 (1.56)
2 1 2 1 2 1 2 8
g1 g
̃
𝛾1 = ̃2 + 1 B2 + 11 g2 g12 + 1 f2 g3 + 63 g2 f 2 − 35 g1 f 3
A2 + f2 B
2 2 72 36 1 18 2 36 2
5 21
− f3 g2 + g1 f2 f3 (1.57)
4 24
g1 g2 g g f
̃
𝛾2 = ̃2 − f2 B2 − f2 A2 − 2 − 2 1 2 + 9 f 3 + 25 f 4 − 53 f3 f 2
B
2 8 8 8 2 18 2 6 2
11 2 7 2 2
− f3 g1 + g1 f2 (1.58)
24 18
and we arrive at the following modified system model
d𝜌
= 𝛼1 𝜌 + (𝛽1 + 𝛽̃1 )𝜌3 + ̃
𝛾1 𝜌5 (1.59)
dt
d𝜃
= (𝛽2 + 𝛽̃2 )𝜌2 + ̃
𝛾2 𝜌4 (1.60)
dt
The approximate solution up to the o(e4 )-th order is
X(t) = 𝜀X1 (t) + 𝜀2 X2 (t) + 𝜀3 X3 (t) + 𝜀4 X4 (t) + o(𝜀5 ),
X1 (t) = 2𝜌(t) cos(−t + 𝜃(t)),
2 2
X2 (t) = − 2f2 𝜌2 (t) + f2 𝜌2 (t) cos(−2t + 2𝜃(t)) + g1 𝜌2 (t) sin(−2t + 2𝜃(t)),
3 3
X3 (t) = 2C1 𝜌3 (t) cos(−3t + 3𝜃(t)) + 2C2 𝜌3 (t) sin(−3t + 3𝜃(t))
X4 (t) = 2D1 𝜌4 (t) cos(−4t + 4𝜃(t)) + 2D2 𝜌4 (t) sin(−4t + 4𝜃(t)). (1.61a)
where
( )
1 5 3 5 31 11
D1 = f2 + f2 f3 − g12 f2 − g1 g2 (1.61b)
15 18 4 9 6
( )
1 43 3 19 3 38
D2 = g1 f2 + f2 g2 + f3 g1 − g13 (1.61c)
15 36 12 2 9
We can calculate a more accurate expression for the limit cycle, and equations
(1.44) and (1.45) become
( )
𝛾1 𝛼1 − 𝛽1 𝛽̃1
̃
𝜌 E = 𝜌E 1 +
M
(1.62)
2𝛽12
10 1 The Asymptotic Perturbation Method for Nonlinear Oscillators
( )2
𝜔M ̃2 + 𝛽̃2 𝜌M
E = 𝜔E + 𝛼 E + 𝛾̃2 𝜌4E (1.63)
If we examine the Eq. (1.59) carefully, we can easily understand that there is the
possibility of two limit cycles and then of a qualitative change in the behavior of the
oscillator with respect to the previsions of the first-order approximation. In fact, we
suppose now that the dissipative coefficient g0 is of order e4 and g2 is chosen in such
a way that g2 − g1 f 2 is of order 𝜀, then 𝛼 1 (see (1.37)) and 𝛽 1 (see (1.38)) are of order
𝜀4 and 𝜀2 , respectively. For example, we can take g0 = 0.0001, g1 = 1, g2 = 1.01, and
f 2 = 1.
Taking into account that 𝜌 must be of order e, then all terms in (39) have the same
magnitude order and we can obtain for the equilibrium values of 𝜌 two positive roots
of order e and then two different limit cycles. Depending on the parameters, the
larger limit cycle and the origin can be asymptotically stable and the smaller unstable
or vice-versa.
By means of the variable change
√
g1 |
|̃ |𝛽 |
| |
t → 2 t, 𝜌 → | 1 |𝜌 (1.64)
𝛽 1 |̃ |
| g1 |
which implies
|̃
𝛾 |𝛼
𝛼1 → 1 2 1 (1.65)
𝛽 1
we can always set (𝛽1 + 𝛽̃1 ) = ±1, ̃ 𝛾1 = ±1.
There are four distinct cases:
(i) (𝛽1 + 𝛽̃1 ) = 1, ̃
𝛾1 = 1: for 𝛼 1 < 0, there is an unstable limit cycle, and for 𝛼 1 > 0,
no limit cycle appears;
(ii) (𝛽1 + 𝛽̃1 ) = −1, ̃𝛾1 = −1: for 𝛼 1 < 0, there is no limit cycle, and for 𝛼 1 > 0, only a
stable limit cycle appears. Note that cases (i) and (ii) are connected by a simple
temporal inversion, followed by the change of the sign of 𝛼 1 .
(iii) (𝛽1 + 𝛽̃1 ) = −1, ̃𝛾1 = 1: for 𝛼 1 < 0, we have a larger stable limit cycle and a
smaller unstable limit cycle; for 𝛼 1 > 0, only the stable limit cycle is present
(Figure 1.1).
(iv) (𝛽1 + 𝛽̃1 ) = 1, ̃𝛾1 = −1: for 𝛼 1 < 0, there is only an unstable limit cycle and for
𝛼 1 > 0, a larger unstable limit cycle and a smaller stable limit cycle are present.
Note that the last cases are also connected by a temporal inversion.
In all these cases, the origin is stable for 𝛼 1 < 0 and unstable for 𝛼 1 > 0.
Amplitude
0.6
0.3
Dissipation coefficient
0.5 0.0 0.5
Figure 1.1 Dissipation (𝛼 1 )–response (𝜌E ) space. Rectangles are stable limit cycles, and
crosses represent unstable limit cycles.
Y
0.10
0.05
–0.05
–0.10
̇
Figure 1.2 Phase space diagram (X(t), Y(t)) with Y(t) = X(t) with f 2 = 2, f 3 = 1, g0 = 0.01,
g1 = 2, and g2 = 2. Circles are the numerical solution, and crosses represent the
approximate solution.
Y 0.4
0.2
–0.2
–0.4
̇
Figure 1.3 Phase space diagram (X(t), Y(t)) with Y(t) = X(t) and f 2 = 2, f 3 = 1, g0 = 0.1,
g1 = 2, and g2 = −2. Circles are the numerical solution and crosses represent the
approximate solution.
1.5 External Excitation in Resonance with the Oscillator 13
and then
d𝜌 f
= 𝛼1 𝜌 + 𝛽1 𝜌3 + sin𝜃 (1.67)
dt 2
dJ f
𝜌 = 𝛽2 𝜌3 + cos 𝜃 (1.68)
dt 2
We can easily determine the equilibrium points and their possible stability. By
means of the variable change
√
1 |𝛼 |
t→ t, 𝜌 → || 1 ||𝜌 (1.69)
|𝛼1 | | 𝛽1 |
which implies
B = 1 + 𝛽22 s = 𝛼1 𝛽1 = ±1 (1.71c)
where
𝛥 = 1 − 3𝛽22 (1.72b)
Amplitude
1.0
0.5
External force
1.0 2.0
Amplitude
2.0
1.0
External force
0.5 1.0
Amplitude
2.0
1.0
External force
0.5 1.0
Figure 1.6 External force–response space. Rectangles are sinks, crosses represent saddle
points, and circles stand for sources.
We now consider the system (1.67)–(1.68) when 𝜌(t) is near (1.44). If the external
excitation is sufficiently small, we obtain
d𝜌 f
= −2𝛼1 (𝜌 − 𝜌2 ) + sin(𝛺t + 𝜃0 ) (1.73)
dt 2
with
𝜃(t) = 𝛺t + 𝜃0 (1.74a)
𝛼2 𝛽1 − 𝛽2 𝛼1
𝛺= (1.74b)
𝛽1
where 𝜃 0 depends on the initial conditions.
Equation (1.85) can be easily resolved and we get
Y
0.20
0.10
–0.10
–0.20
Figure 1.7 Associated map of the nonautonomous Eq. (1.27) with f 2 = −1, f 3 = −1,
g0 = 0.02, g1 = 1, g2 = −3, and f = 3.5 ⋅ 10−4 . Crosses are the approximate solution, and
circles represent the numerical solution.
1.6 Conclusion
We have demonstrated the power of the AP method and how it produces useful
approximate solutions. In particular, we have treated a class of strongly nonlinear
oscillators subject to an external periodic force in resonance with the natural fre-
quency of the oscillator.
We have found bifurcations and limit cycles, which are influenced by the presence
of external excitation. In addition, we can observe a quasiperiodic motion, charac-
terized by the combination of the natural frequency with a low frequency connected
to the external excitation.
1.6 Conclusion 17
2.1 Introduction
A parametrically excited Liénard system is investigated by an asymptotic per-
turbation method based on Fourier expansion and time rescaling. Two coupled
equations for the amplitude and the phase of solutions are derived. Their fixed
points correspond to limit cycles for the Liénard system, and we determine the
stability of steady-state response as well as response–parametric excitation and
response–frequency curves. We use the Poincarè–Bendixson theorem, the Dulac’s
criterion, and energy considerations to study the existence and characteristics of
limit cycles of the two coupled equations. A limit cycle corresponds to a mod-
ulated motion in the Liénard system. We show that modulated motion can also
be obtained for very low values of the parametric excitation and construct an
approximate analytic solution. Moreover, we observe an unusual infinite-period
homoclinic bifurcation, as in certain cases, due to the symmetry of the two coupled
equations, two stable limit cycles approach a saddle point and merge to form a
greater stable limit cycle. Subsequently, this limit cycle and another unstable limit
cycle coalesce and annihilate through a saddle–node bifurcation. The comparison
with the solution obtained by the numerical integration confirms the validity of our
analysis. It is well known that many mechanical systems and oscillating circuits
can be modeled by the Liénard system [65, 84, 226]
̈ + f (X(t)) = g(X(t))X(t)
X(t) ̇ (2.1)
where the dot denotes differentiation with respect to the time and the arbitrary func-
tions f (x) and g(x) are supposed to be analytic. Obviously, Eq. (2.1) can be interpreted
mechanically as the equation of motion for a unit mass subject to a nonlinear damp-
ing force g(x)ẋ and a nonlinear restoring force −f (x).
We restrict our study to the following particular case of Eq. (2.1)
X(t) ̇
̈ + 𝜔20 X(t) + f2 X 2 (t) + f3 X 3 (t) = g0 X(t) ̇
+ g1 X(t)X(t) ̇
+ g2 X 2 (t)X(t) (2.2a)
corresponding to the choice
f(x) = 𝜔20 x + f2 x2 + f3 x3 , g(x) = g0 + g1 x + g2 x2 (2.2b)
Equation (2.2a,b) can be also considered a generalized Van der Pol–Duffing sys-
tem, because it includes as particular cases the Van der Pol oscillator ( f 2 , f 3 , g1 = 0
and g0 = −g2 ≠ 0) and the Duffing equation (f 2 = g1 = g2 = 0 and g0 , f 3 ≠ 0). Through
a second-order perturbation analysis, we have investigated Eq. (2.2a,b) in [119], and
the existence of one or two limit cycles has been demonstrated. Moreover, a resonant
external excitation has been added to Eq. (2.2a,b) and, by a lower order perturbation
analysis, we have derived a sufficient condition to obtain a two-period quasiperi-
odic motion when a second low frequency appears in addition to the forcing fre-
quency. Finally, analytic approximate solutions have been checked by numerical
integration.
In this chapter, we consider the Liénard oscillator in the case of parametric exci-
tation. In particular, we introduce a principal parametric excitation 2fX(t) cos(2𝛺t)
into Eq. (2.2a,b) and obtain
X(t) ̇
̈ + 𝜔20 X(t) + f2 X 2 (t) + f3 X 3 (t) = g0 X(t) ̇
+ g1 X(t)X(t)
2 ̇
+ g2 X (t)X(t) + 2fX(t) cos(2𝛺t) (2.3)
Principal parametric resonance or 1/2-subharmonic external resonance has been
studied in many papers and systems similar to Eq. (2.3) have been considered in
[13, 204].
In Section 2.2, we use the AP method to calculate an approximate analytic
solution. We derive a model system of two coupled differential equations in the
phase and amplitude of solutions. Stable (unstable) fixed points of the model system
correspond to stable (unstable) limit cycles of the parametrically excited Liénard
system (2.3). Subsequently, we compare frequency–response and parametric
excitation–response curves with the results of the numerical integration.
In Section 2.3, we perform a global analysis of the model system by the
Poincarè–Bendixson theorem, the Dulac’s criterion, and energy considerations.
We derive several conditions in order to exclude or permit the existence of limit
cycles, which correspond to the modulated motions for the Liénard systems (2.3).
Moreover, we study the modulated motion that appears for very low values of the
parametric excitation. In this case, the two-period quasiperiodic motion is charac-
terized by a slight modulation of the fundamental oscillation with a modulation
amplitude proportional to the magnitude of the parametric excitation.
In Section 2.4, we demonstrate the presence of an unusual infinite-period homo-
clinic bifurcation. In certain cases, the symmetry of the model system gives rise to
the appearance of two stable limit cycles, a saddle point, and a surrounding unstable
limit cycle. If we increase the parametric excitation, the two limit cycles approach
the saddle point and subsequently merge to form a greater stable limit cycle. Finally,
with a further increase in the parametric excitation, this limit cycle and the unstable
limit cycle coalesce and disappear through a saddle–node bifurcation.
Note that in the usual homoclinic bifurcation, a limit cycle approaches a saddle
point, and at the bifurcation point, the limit cycle and a branch of both the stable and
unstable manifolds of the saddle point coincide, forming a homoclinic connection,
which is essentially a cycle limit with an infinite period. As the control parame-
ter increases, the limit cycle suddenly vanishes (“blue sky catastrophe”). On the
2.2 Periodic Solutions and Their Stability 21
with 𝛾 n = |n| for n ≠ 0, and 𝛾 0 = r is a positive number, which will be fixed later on;
being the assumed solution X(t) is real, we have
𝜓n (t, 𝜀) = 𝜓−n
∗
(t, 𝜀) (2.7)
The assumed solution (2.6) can be written more explicitly
X(t) = 𝜀r 𝜓0 (𝜏; 𝜀) + 𝜀𝜓1 (𝜏; 𝜀) exp(−i𝛺t) + 𝜀2 𝜓2 (𝜏; 𝜀) exp(−2𝛺it) + O(𝜀3 ) (2.8)
22 2 The Asymptotic Perturbation Method for Remarkable Nonlinear Systems
and we see that it can be considered a combination of the various harmonics with
coefficients depending on 𝜏 and 𝜀.
We suppose that the functions 𝜓 n (𝜏, 𝜀)s can be expanded in power series of, i.e.
∑
∞
𝜓n (𝜏; 𝜀) = 𝜀i 𝜓n(i) (𝜏) (2.9)
i=0
We have assumed in Eq. (2.9) that the limit of the 𝜓 n (𝜏; 𝜀)’s for e → 0 exists and is
finite. In the following, for simplicity, we use the abbreviations 𝜓n(0) = 𝜓n for n ≠ 1
and 𝜓1(0) = 𝜓 for n = 1.
Note that the introduction of the slow time (2.5) implies that
( )
d q d𝜓n
(𝜓 exp(−in𝛺t)) = −in𝛺𝜓n + 𝜀 exp(−in𝛺t) (2.10)
dt n d𝜏
In order to determine the coefficients 𝜓 n (𝜏; 𝜀), we insert the assumed solution
(2.8) into Eq. (2.4), and obtain various equations for each harmonic n and for a fixed
order of approximation on the perturbation parameter 𝜀.
For n = 0, we get
2f2
𝜀r 𝜓0 + |𝜓|2 𝜀2 + O(𝜀4 , 𝜀r+2 ) = 0 (2.11)
𝛺2
A correct balance of terms shows that r = 2 and then we derive the following rela-
tion
2f
𝜀2 𝜓0 = −𝜀2 22 (|𝜓|2 ) + O(𝜀4 ) (2.12)
𝛺
For n = 2, taking into account Eq. (2.10), we have
and then
f2 + ig1 𝛺 2 2
𝜀2 𝜓2 = 𝜀 𝜓 + O(𝜀4 ) (2.14)
3𝛺2
For n = 1, Eq. (2.4) yields
d𝜓 1+q
2i𝛺 𝜀 − 2𝜀3 𝜎𝛺𝜓 − 2𝜀3 f2 (𝜓0 𝜓 + 𝜓2 𝜓 ∗ ) − 3𝜀3 f3 |𝜓|2 𝜓 − i𝜀3 𝛺g0 𝜓
d𝜏
− i𝜀3 g1 𝛺(𝜓0 𝜓 + 𝜓2 𝜓 ∗ ) − i𝜀3 g2 𝛺|𝜓|2 𝜓 + f 𝜓 ∗ + O(𝜀5 , 𝜀1+2q ) = 0 (2.15)
If q = 2, the first term has the same magnitude order of all the other terms.
As we can see from Eqs. (2.12) and (2.14), we can derive a differential equation for
the evolution of the complex amplitude,
d𝜓
= (𝛼1 + i𝛼2 )𝜓 + (𝛽1 + i𝛽2 )|𝜓|2 𝜓 + i𝛾𝜓 ∗ (2.16)
d𝜏
with
g0 f
𝛼1 = , 𝛼 = −𝜎, g = (2.17)
2 2 2𝛺
( )
g gf 1 g21 3f3 5f 2
𝛽1 = 2 − 1 22 , 𝛽2 = − + (2.18)
2 2𝛺 𝛺 6 2 3𝛺2
2.2 Periodic Solutions and Their Stability 23
where
( )
𝛥 = 𝛾 2 1 + 𝛽22 − (𝛼1 𝛽2 − 𝛼2 𝛽1 )2 (2.27)
0.2
0.18
C
0.16
0.14
0.12
B
0.1
0.08 A
0.06
0.04
0.02
Figure 2.1 Response–parametric excitation curve, i.e. amplitude of the response (𝜌) as
function of the excitation (f ). Solid lines stand for stable and dashed lines for unstable
solutions, while boxes correspond to the numerical exact solution.
0.2
ρ
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
Figure 2.2 Frequency–response curve, i.e. the response (𝜌) as function of the detuning (𝜎).
Solid lines stand for stable and dashed lines for unstable solutions.
is possible, which is stable. If 𝜎 A < 𝜎 < 𝜎 B , there are two possible solutions: a nontriv-
ial solution, which is stable, and a trivial solution, which is unstable. If 𝜎 > 𝜎 B , there
are three possible solutions, the trivial solution, which is stable; and two nontrivial
solutions, the larger of which is stable. The subcritical instability appears because,
in the latter case, the response depends on the initial conditions and the system
tends asymptotically to the trivial solution or to the steady-state periodic response.
If 𝜎 > 𝜎 C , only the stable trivial solution is possible.
In order to verify if the AP method is a valid tool for approximate solutions of
nonlinear oscillators with small dissipation coefficients, we compare the analytic
approximate solution with the numerical integration in the following two examples.
We have chosen the following set of parameters
f2 = 1.0, f3 = −0.11, g0 = −0.01, g1 = 1.0, g2 = −1.0, f = 0.01, 𝜎 = 0.01 (2.31)
2.3 Global Analysis of the Model System 27
Y
0.2
0.1
–0.1
–0.2
Figure 2.3 Phase-space diagram (X(t), Y(t)) with Y(t) = X(t) ̇ with f 2 = 1.0, f 3 = −0.11,
g0 = −0.01, g1 = 1.0, g2 = −1.0, s = 0.01, and f = 0.01. Circles are the numerical solution, and
crosses represent the approximate solution.
We expect the appearance of a stable limit cycle with amplitude 𝜌0 = 0.085 (see
̇
Eq. (2.25)). In Figure 2.3, we show in the phase space (X(t), Y (t) = X(t)), a compar-
ison between the analytic approximate solution (2.22) and the numerical solution:
circles represent the numerical solution and crosses represent the approximate solu-
tion. As the solution repeats itself one cycle after another, we have represented only
a cycle. The agreement of the results appears to be excellent, because the maxi-
mum difference is 3 × 10−3 and the medium difference is 1 × 10−3 , i.e. of order e3
as expected.
In Figure 2.4, we have increased the parametric excitation by an order of mag-
nitude (f = 0.1). According to the model system (2.20) and (2.21) and Eq. (2.25), in
this case, we have a stable limit cycle with amplitude 𝜌0 = 0.165. The maximum dif-
ference between the approximate solution (2.22) and the numerical solution is now
0.04 and the medium difference is 0.02, i.e. of order e3 as expected.
In this section, we perform a global analysis of the model system (2.20) and (2.21)
and determine several conditions for the existence of limit cycles. They correspond
to two-period quasiperiodic solutions of the Liénard system (2.4).
Another random document with
no related content on Scribd:
TOM.
I now come to the very prince of pets, the one of all I ever had the
most noble and most dear,—Tom, a Newfoundland setter, the
favorite dog of my brother Albert. He has been a member of our
family for five or six years past. We brought him from the city to our
pleasant village home in Pennsylvania, where we now live.
Tom is a dog of extraordinary beauty, sagacity, and good feeling. He
is very large, and, with the exception of his feet and breast, jet black,
with a thick coat of fine hair, which lies in short curls, glossy and
silken. He has a well-formed head, and a handsome, dark eye, full of
kindness and intelligence. His limbs are small, and his feet
particularly delicate. He is, I am sorry to say, rather indolent in his
habits, always prefers to take a carriage to the hunting-ground, when
he goes sporting with his master, and he sleeps rather too soundly at
night to be a good watch-dog. We make him useful in various ways,
however, such as carrying baskets and bundles, and sometimes we
send him to the post-office with and for letters and papers. These he
always takes the most faithful care of, never allowing any one to look
at them on the way. He is a remarkably gentlemanly dog in his
manner, never making free with people, or seeming too fond at first
sight; but if you speak to him pleasantly, he will offer you a friendly
paw in a quiet way, and seem happy to make your acquaintance. He
never fawns, nor whines, nor skulks about, but is dignified, easy, and
perfectly at home in polite society. He is a sad aristocrat, treats all
well-dressed comers most courteously, but with shabby people he
will have nothing to do. Tom knows how to take and carry on a joke. I
recollect one evening, when we had visitors, and he was in the
parlour, I put on him a gay-colored sack of my own, and a large
gypsy hat, which I tied under his throat. Instead of looking ashamed
and trying to get these off, as most dogs would have done, he
crossed the room and sprang on to the sofa, where he sat upright,
looking very wise and grave, like some old colored woman at church.
The illustrious General Tom Thumb once travelled with my brother
and this dog, and, falling very much in love with his namesake,
offered any price for him. Of course, my brother would not think for a
moment of selling his faithful friend, and even had he felt differently, I
doubt very much whether Tom, who had been used to looking up to
full-grown men, would have shown much obedience or respect, for
such a funny little fellow as the General. It was amusing to observe
the dog’s manner toward his small, new acquaintance. He was kind
and condescending, though he sometimes seemed to think that the
General was a little too much inclined to take liberties with his
superiors in age and size,—rather more forward and familiar than
was quite becoming in a child.
Two or three years ago, Tom was the beloved playfellow of my
brother Frederic’s youngest daughter,—our little Jane. She always
seemed to me like a fairy-child, she was so small and delicate, with
such bright golden curls falling about her face,—the sweetest face in
the world. It was beautiful to see her at play with that great, black
dog, who was very tender with her, for he seemed to know that she
was not strong. One evening she left her play earlier than usual, and
went and laid her head in her mother’s lap, and said, “Little Jane is
tired.” That night she sickened, and in a few, a very few days she
died. When she was hid away in the grave, we grieved deeply that
we should see her face no more, but we had joy to know that it
would never be pale with sickness in that heavenly home to which
she had gone; and though we miss her still, we have great
happiness in the thought that she will never be “tired” any more, for
we believe her to be resting on the bosom of the Lord Jesus.
One day last spring, I remember, her mother gave me a bunch of
violets, saying, “They are from the grave of little Jane.” I suppose
they were like all other blue violets, but I thought then I had never
seen any so beautiful. It seemed to me that the sweet looks of the
child were blooming out of the flowers which had sprung up over the
place where we had laid her.
Tom seems much attached to all our family, but most devotedly so to
my brother Albert. They two have hunted very much together, and
seem equally fond of the sport. If Tom sees his master with his
hunting-dress on, and his fowling-piece in hand, he is half beside
himself with joy. But when he returns from the hunt, spent and weary,
he always comes to me to be fed and petted.
You will remember that years have passed by since this brother and I
were schoolmates and playmates together. He is now a fine young
man, while I am a full-grown woman, who have seen the world I
used to think so grand and glorious, and found it—no better than it
should be. But of my brother. He is our youngest, you know, and so
has never outgrown that peculiar fondness, that dear love, we
always give to “the baby.” While I have been writing these histories,
and recalling in almost every scene the playmate of my childhood, I
can only see him as a boy,—a little black-eyed, rosy-cheeked boy; it
is very difficult to think of him as a man, making his own way bravely
in the world. Last spring we observed that dear Albert’s bright face
had become very thoughtful and serious; we knew that something
was weighing on his mind, and finally it came out. He was about to
leave us all for a long time, it might be for ever; he was going to
California! We were very unhappy to hear this, but, as it was on
some accounts the best thing that my brother could do, we finally
consented, and all went to work as cheerfully as we could to help
him off.
It was a bright May morning when he left, but it seemed to us that
there never was a darker or sadder day. The dear fellow kept up
good courage till it came to the parting; then his heart seemed to
melt and flow out in his tears, fast dropping on the brows and necks
of his mother and sisters, as he held them for the last time to his
heaving breast. But I will not dwell on this parting, for my own eyes
grow so dim I cannot well see to write.
I remember that poor Tom seemed greatly troubled that morning; he
knew that something sad was happening, and looked anxiously in
our faces, as though he would ask what it was; and when my brother
patted him on the head, bade him good by, and passed out of the
gate, forbidding him to follow, the faithful creature whined sadly, and
looked after him wistfully, till he was out of sight.
After Albert had been gone about an hour, I remember that I went up
into his room, and sat down in his favorite seat, by the window. O,
how still and lonely and mournful it seemed there! Near me hung my
brother’s fencing-sword and mask, which he had used only the day
before,—on the floor lay the game-bag, which he had always worn in
hunting, and which he had flung out of his trunk, not having room for
it. This brought my merry brother before me more clearly than any
thing else. I took it up and held it a long time, mourning at heart, but I
could not weep. Suddenly I heard a low whine in the hall, and Tom
stole softly into the room. He came to me and laid his head in my
lap; but when he saw the game-bag there, he set up a most mournful
cry. Then I flung my arms about him, bowed my head down against
his neck, and burst into tears. I forgot that he was a poor dumb
brute, and only remembered that he loved my brother, and my
brother loved him, and that he mourned with me in my sorrow. After
this, it was very affecting to see Tom go every day, for a long while,
to the gate, out of which he had seen his master pass for the last
time, and then stand and look up the street, crying like a grieved
child.
As you will readily believe, Tom is now dearer than ever to us all; we
cannot see him without a sweet, sad thought of that beloved one so
far away. I am not now at home, but I never hear from there without
hearing of the welfare of the noble dog which my brother, in going,
bestowed upon me.
SUPPLEMENTARY STORIES.
It is twenty years since the first part of this little volume was
published. The dear children for whom those simple stories of my
childhood were told are men and women now, and wonderful
changes have taken place in all our lives and in all the world. But in
growing old I have not lost any thing of my old love of pets; and I
hope that my little readers of this time will understand and share that
feeling. I hope that you, dear boys and girls, look on all innocent
dumb creatures about you as friends, and have not only a kindly
interest in them, but respect them for all that is lovely and wonderful
in their brief existences, and as objects of the unceasing care and
tenderness of our Father in heaven. Every smallest creature that
lives represents a thought of God,—was born out of his great, deep,
infinite life.
I hope you especially like to hear about dogs and cats, birds and
chickens, for it is of them that I have a few new stories to relate, as
true as they are amusing or marvellous.
FIDO THE BRAVE.
FAITHFUL GRIMALKIN.
Many years ago, when my parents lived in old Connecticut, my
mother had a pet cat, a pretty graceful creature, frisky and arch and
gay, though clad in sober gray. She was a favorite with all the large
household, but especially attached herself to my mother, following
her about everywhere,—“up stairs, down stairs, and in my lady’s
chamber,” accompanying her in her walks, hiding behind every bush,
and prancing out upon her in a surprising, not to say startling,
manner.
At last she grew out of kittenhood, laid aside, in a measure, kittenish
things, and became the happiest, fondest, proudest feline mamma
ever beheld. She caressed and gloated over her little, blind, toddling,
mewing, miniature tigers in a perfect ecstasy of maternal delight.
Just at this interesting period of pussy’s life our family moved from
the old place to a house in the country, about a mile away. My
mother was ill, and was carried very carefully on a bed from one
sick-room to another. In the hurry, trouble, and confusion of that time,
poor pussy, who lodged with her family in an attic, was quite
forgotten. But early in the morning of the first day in the new house,
—a pleasant summer morning, when all the doors and windows
were open,—as my mother lay on her bed, in a parlor on the first
floor, she saw her cat walk into the hall and look eagerly around. The
moment the faithful creature caught sight of her beloved mistress,
she came bounding into the room, across it, and on to the bed,
where she purred and mewed in a delighted, yet reproachful way,
quite hysterical, licking my mother’s hand and rubbing up against her
cheek in a manner that said more plainly than words, “Ah! my dear
madam, didst thou think to leave thy faithful Grimalkin behind?
Where thou goest, I will go.”
She was taken into the kitchen and treated to a cup of new milk; but
after a few moments given to rest and refreshment she disappeared.
Yet she went only to come again in the course of an hour, lugging
one of her kittens, which she deposited on the bed, commended to
my mother’s care, and straightway departed. In an almost incredibly
short time she came bounding in with a second kitten. She continued
her journeys till the whole litter had been safely transported, over hill
and dale, ditches and stone-walls, through perils of unfriendly dogs
and mischievous boys, and the family flitting was complete.
After this, our noble puss was loved and respected more than ever.
She dwelt long in the land, and her kits grew up, I believe, to be
worthy of such a mother.
OBEDIENT THOMAS.
Now I want to give you an instance of filial respect and submission
in a young cat. When we first came to Washington, nearly two years
ago, I took to petting a handsome cat belonging to the relatives with
whom we then lived. I fed and caressed her, and she became very
fond of me, always running to meet me when I entered the garden
which she haunted, or the barn in which she lodged. She was rather
wild in her ways, and so stole a nest, in which she finally hid away
some kittens, that she afterwards reared to be wilder than herself.
These somehow disappeared, all but one, which, when he was
about half grown, I undertook to tame. It was a difficult, tedious job;
but I persevered, and at last found him a more affectionate, docile
pet than ever his mother had been. She had seemed fond of him in
his wild, unregenerate days, but as soon as he became
domesticated, and I began to show a partiality for him, she grew very
severe with him, scratching his face and boxing his ears whenever
she saw me caressing him. I soon noticed that when she was near
he was shy, pretending not to be on intimate terms with me; while, if
she was out of the way, I had only to call his name, to have him
come galloping up from the furthest part of the long garden, to rub
against me, to lick my hand, and show every feline fondness and
delight. Now we live at another house, and I seldom see my pets,
mother and son; but they are loving and constant still, proving that
the poet Coleridge didn’t know every thing when he talked about “the
little short memories” of cats.
Master Thomas has grown large and strong, and is accounted a
gallant young fellow by all the young pussies in the neighborhood.
But while toward cats of his own sex he is fierce and combative, he
is just as meek and deferential to his mother as he was in his tender
kittenhood. The other day I encountered him in the old garden, and
was surprised to find how stalwart he had become. I stooped to
caress him, and he seemed as susceptible to gentle overtures as
ever, arched his back, switched his tail, and purred rapturously.
Suddenly the mother cat stole out from behind a tree, and confronted
us. “Good morning, madam,” I said, for I always talk to cats and dogs
just as I talk to other people. “You have a fine son here; a handsome
young fellow, that favors you, I think.” But she wasn’t to be softened
by the compliment. She walked straight up to him, and boxed him
first on one ear and then on the other, quite in the old motherly way.
As for him he never thought of resenting the old lady’s act, or
opposing her will, but drooped his lordly tail, and hastily retreated.
Now that is what I call good family discipline.
This city of Washington is a place where the wits of people are
sharpened, if anywhere, and perhaps even cats and dogs become
uncommonly clever and knowing here. Only yesterday I was told of a
Washington cat which had just been found out in a wonderful trick.
Observing that, when the door-bell rang, the one servant of the
household was obliged to leave the kitchen, she managed to slyly
ring the bell, by jumping up against the wire, and invariably, when
her enemy, the cook, went to the door, she would slip into the
kitchen, and help herself to whatever tempting article of food was
within reach. At last some one watched, and caught her at her secret
“wire-pulling.” Poor puss retired with a drooping tail and a most
dejected aspect, evidently realizing that the game was up.
Another cat I know of was of so amiable and benevolent a
disposition that she actually adopted into her own circle of infant kits
a poor, forlorn little foundling of a rat. As her nursling he grew and
thrived, seeming quite as tame as the others; and when a
mischievous boy set a rat-terrier on him, and so finished him, cat and
kittens really seemed to mourn for their foster son and brother.