Density of instantaneous frequencies in the Kuramoto-Sakaguchi model Julio D. da Fonseca1 Edson D. Leonel1 and Rene O. Medrano-T1 2_2

2025-05-06 0 0 1.34MB 29 页 10玖币
侵权投诉
Density of instantaneous frequencies in the
Kuramoto-Sakaguchi model
Julio D. da Fonseca1
, Edson D. Leonel1
, and Rene O. Medrano-T1 2
December 13, 2022
1Departamento de Física, Universidade Estadual Paulista, Bela Vista, 13506-900 Rio Claro,
SP, Brazil
2Departamento de Física, Universidade Federal de São Paulo, UNIFESP, 09913-030, Campus
Diadema, São Paulo, Brasil
Abstract
We obtain a formula for the statistical distribution of instantaneous frequencies in the
Kuramoto-Sakaguchi model. This work is based on the Kuramoto-Sakaguchi’s theory of
globally coupled phase oscillators, which we review in full detail by discussing its assump-
tions and showing all steps behind the derivation of its main results. Our formula is a
stationary probability density function with a complex mathematical structure, is consis-
tent with numerical simulations and gives a description of the stationary collective states
of the Kuramoto-Sakaguchi model.
1 Introduction
Synchronization is the process by which interacting oscillatory systems adjust their frequencies
in order to display the same common value [1]. Power grids [2], semiconductor laser arrays
[3], cardiac pacemaker cells [4], and neurosciences [5] are just a few examples in a multitude of
domains where synchronization is an active research subject.
jcddafonseca@gmail.com
edson-denis.leonel@unesp.br
rene.medrano@unifesp.br
1
arXiv:2210.05011v4 [nlin.AO] 12 Dec 2022
The works of A. Winfree and Y. Kuramoto brought seminal contributions to the study of
synchronization [4, 6, 7, 8]. Inspired by Winfree’s pioneering ideas [9], Kuramoto formulated a
model of coupled phase oscillators today known as Kuramoto model. The Kuramoto model was
introduced in Ref. [7], and its first and more detailed analysis by Kuramoto himself, published
in Ref. [8]. Since then, many studies about the Kuramoto model and its variants appeared
in the literature. (Reviews about the Kuramoto model can be found in Refs. [10, 11]; see
Refs. [12, 13, 14, 15] for later studies related to variants of the Kuramoto model and their
applications.)
The Kuramoto model consists of an ensemble of oscillators with a mean-field coupling and
randomly distributed natural (or intrinsic) frequencies. An oscillator is characterized by its
phase, and the first-order time derivative of the oscillator’s phase, which here we call instan-
taneous frequency, is defined by an autonomous first-order ordinary differential equation. The
theoretical analysis of the Kuramoto model [8, 11] evinces a transition between two station-
ary collective states: an incoherent state and a synchronization one. In the incoherent state,
instantaneous and natural frequencies have the same statistical distribution. In the synchro-
nization state, some oscillators have instantaneous frequencies sharing the same value. The
number of synchronized oscillators depends on the model’s parameter called coupling-strength,
and synchronization only occurs for a coupling strength above a critical value [8, 11]. In a
simplified version of the Kuramoto model, identical (with the same natural frequencies) and
symmetrically-coupled oscillators show multiple regular attractors [16], and the synchronization
state is the most probable one[17, 18].
H. Sakaguchi and Y. Kuramoto created a generalization of the Kuramoto model [19] intro-
ducing into the coupling function a phase shift, also called phase-lag. The Kuramoto-Sakaguchi
model and its variants appear in the study of a wide range subjects such as chimera states
[20, 21], chaotic transients [22], pulse-coupled oscillators [23], and Josephson-junction arrays
[24]. In addition, the coupling function with a phase-lag can be interpreted as an approximate
model of interactions with time-delayed phases [25]. The Kuramoto-Sakaguchi model exhibits
the same stationary collective states as the original Kuramoto model [19].
Collective states of Kuramoto-like models are commonly characterized by means of an order
parameter, which is zero in the incoherent state and takes finite values in the synchronization
state. In this work, we follow a different approach from the usual order-parameter analysis
uncovering how instantaneous frequencies are statistically distributed in the stationary collec-
tive states of the Kuramoto-Sakaguchi model. Instantaneous frequencies collectively reflect
the occurence of synchronized behavior [26], and they are also relevant in the study of other
phenomena (e.g. frequency spirals [32]).
We will show how to obtain a formula for the statistical distribution of instantaneous fre-
quencies. The formula is defined by a stationary probability density function, which we refer
to as density of instantaneous frequencies. Our goal is similar to the one pursued in Ref. [26]
for the Kuramoto model, but here we will show how to obtain a more general result in a more
straightforward way. A related (but still rather a different) problem was addressed by Sakaguchi
2
and Kuramoto in Ref. [19], where they analyzed the statistical distribution of coupling-modified
frequencies, namely instantaneous frequencies averaged over infinitely large time intervals (see
Refs. [11, 26] for further details).
This work is based on the Kuramoto-Sakaguchi theory, described, as far as we know, only
in Ref. [19]. We will discuss the fundamental assumptions of the Kuramoto-Sakaguchi theory
and detail how its main results can be derived. Our opinion is that an explicit presentation of
the Kuramoto-Sakaguchi theory is still absent.
We organized this paper as follows. In Section 2, we present the Kuramoto-Sakaguchi the-
ory and state diagrams pointing out the transition between the incoherent and synchronization
states. In Section 3, we extend the Kuramoto-Sakaguchi theory by providing additional ana-
lytical results and obtaining the formula of the density of instantaneous frequencies. In Section
4, we discuss the properties of our formula in a specific application example, in which natural
frequencies have a Gaussian statistical distribution. In Section 5, we check the consistency of
our formula with numerical simulation data. Conclusions and an outlook on possible research
directions are given in Section 6.
2 Kuramoto-Sakaguchi theory
The Kuramoto-Sakaguchi (KS) model[19] consists of an infinitely large number Nof all-to-all
coupled oscillators. The state of an oscillator of index i= 1...N is characterized by its phase
θi, which changes in time according to
˙
θi=ωiK
N
N
X
j=1
sin(θiθj+α),(1)
where ˙
θiis the first-order time-derivative of θi,ωiis a random number with a prescribed density1,
and Kand αare real constant parameters. We refer to ˙
θias the instantaneous frequency, ωi
is called natural frequency and K, the coupling strength. The oscillator of index jcan be
represented by the complex number exp(j). Oscillators are then points in a complex plane
moving over a unit-radius circle centered at the origin.
A valuable concept for the analysis of collective behavior in the KS model is that of mean
field, defined by
Z=1
N
N
X
j=1
exp(j),(2)
1For the sake of simplicity, hereafter we always use the term “density” to refer to a probability density
function.
3
which can be interpreted as the average oscillator state. The mean field can be written as a
complex number
Z=Rexp(iΘ) (3)
where Θdenotes the mean-field phase and R, the mean-field modulus, referred to as order
parameter. If the oscillators are quasi-aligned, i.e., they have approximately the same phase,
then R'1. Yet, a quasi-uniform scattering of all oscillator-points over the circle results in a
mean-field located near the origin, i.e. R'0.
For N→ ∞, the mean field, at a time instant t, is given by
Z=
+π
ˆ
π
exp()n(θ, t), (4)
where n(θ, t)is the density of phases at the same time instant. Two simple scenarios are
assumed concerning the properties of n(θ, t)in the long-time (t→ ∞) and large-size (N
) limits. First, n(θ, t) = 1
2πfor π < θ +π, and n(θ, t)=0, otherwise, i.e. n(θ, t)is a
time-independent and uniform density (the value 1
2πcomes from the normalization condition).
Second, n(θ, t)is a steadily traveling wave with velocity , i.e. n(θΩ∆t, t) = n(θ, t + ∆t)for
any time instant tand time interval t. This means that the wave profile does not change in
time, and the wave propagates with constant-in-time velocity . We call the wave-propagation
velocity, ,synchronization frequency.
The first scenario defines the incoherent state, and the second, the synchronization state. In
the incoherent state, oscillators are uniformly spread over the unit circle. In the synchronization
state, a bunch of oscillators is synchronized, that is, they change collectively their phase at the
same constant rate .
After inserting a uniform phase density in Eq. (4), we see that Z= 0. So, from Eq.
(3), the order parameter (R) is zero in the incoherent state. Yet, if a KS system exhibits
synchronization, then the assumption of a traveling wave with a stationary and non-uniform
profile, moving with constant velocity , means that Ris finite and time-independent. Moreover,
Zmoves in the complex plane following a circular path of radius Rand velocity . That is, a
uniform circular motion given by
Z(t) = Rexp [i(Ωt+ Θ0)] ,(5)
where Θ0is the mean-field phase at an arbitrary initial time instant.
Let us consider KS oscillators in a different complex plane, with the same origin as the
previous one, but with both axis rotating with angular velocity . In the new rotating frame, we
represent the oscillator of index jby the complex number exp(j), where ψjis the oscillator’s
phase. The analogous of Eqs. (2), (3), and (4) are
Z0=1
N
N
X
j=1
exp(j),(6)
4
Z0=Rexp(iΨ),(7)
and, for N→ ∞,
Z0=
+π
ˆ
π
exp()n(ψ). (8)
The quantities Z0and n(ψ)are representations of the mean field and the phase density in
the rotating frame. Comparing Eq. (3) to Eq. (7), we see that Zand Z0have the same length
R. The mean-field length is invariant to the change of frames because the phase density profile
is kept unchanged. Note also that, in Eq. (8), we removed the time dependence from the phase
density, since both the rotating frame and the steadily traveling wave move together with the
same velocity . So, both Rand Ψare time-independent, which is the same as stating that
the mean field is fixed in the rotating frame.
Some conventions are useful to simplify the analysis of the KS model at a time instant
t > to= 0, with t0denoting the initial time instant. We choose a fixed frame such that its real
axis has the same direction as the mean field at time t0. So, Θ0= 0 and Z(0) = R. Another
important convention is defining a rotating frame such that, at the initial time t0, its real axis
is dephased by αfrom the fixed-frame’s real axis (the same parameter αof Eq. (1)). This is
the same as setting Ψ = α.
Thus, from Eqs. (5) and (7), at the time instant t,
Z(t) = Rexp (it)(9)
and
Z0=Rexp iα. (10)
Also, as a consequence of the above conventions, a simple geometric inquiring yields the relations
˙
ψi=˙
θi(11)
and
ψi=θit+α. (12)
In Eq. (11), ˙
ψiis the instantaneous frequency of an oscillator of index iin the rotating frame.
Using Eqs. (11) and (12), we can recast Eq. (1) as
˙
ψi=ωiK
N
N
X
j=1
sin(ψiψj+α).(13)
Multiplying the right-hand sides of Eqs. (6) and (10) by exp [i(ψi+α)] and equating their
imaginary parts result in
N1
N
X
j=1
sin(ψiψj+α) = Rsin ψi.(14)
5
摘要:

DensityofinstantaneousfrequenciesintheKuramoto-SakaguchimodelJulioD.daFonseca1*,EdsonD.Leonel1„,andReneO.Medrano-T12…December13,20221DepartamentodeFísica,UniversidadeEstadualPaulista,BelaVista,13506-900RioClaro,SP,Brazil2DepartamentodeFísica,UniversidadeFederaldeSãoPaulo,UNIFESP,09913-030,CampusDiad...

展开>> 收起<<
Density of instantaneous frequencies in the Kuramoto-Sakaguchi model Julio D. da Fonseca1 Edson D. Leonel1 and Rene O. Medrano-T1 2_2.pdf

共29页,预览5页

还剩页未读, 继续阅读

声明:本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知玖贝云文库,我们立即给予删除!
分类:图书资源 价格:10玖币 属性:29 页 大小:1.34MB 格式:PDF 时间:2025-05-06

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 29
客服
关注