A manifold independent approach to understanding transport in stochastic dynamical systems

Erik M. Bollt, Lora Billings, Ira B. Schwartz

Research output: Contribution to journalArticle

28 Citations (Scopus)

Abstract

We develop a new collection of tools aimed at studying stochastically perturbed dynamical systems. Specifically, in the setting of bi-stability, that is a two-attractor system, it has previously been numerically observed that a small noise volume is sufficient to destroy would be zero-noise case barriers in the phase space (pseudo-barriers), thus creating a pre-heteroclinic tangency chaos-like behavior. The stochastic dynamical system has a corresponding Frobenius-Perron operator with a stochastic kernel, which describes how densities of initial conditions move under the noisy map. Thus in studying the action of the Frobenius-Perron operator, we learn about the transport of the map; we have employed a Galerkin-Ulam-like method to project the Frobenius-Perron operator onto a discrete basis set of characteristic functions to highlight this action localized in specified regions of the phase space. Graph theoretic methods allow us to re-order the resulting finite dimensional Markov operator approximation so as to highlight the regions of the original phase space which are particularly active pseudo-barriers of the stochastic dynamics. Our toolbox allows us to find: (1) regions of high activity of transport, (2) flux across pseudo-barriers, and also (3) expected time of escape from pseudo-basins. Some of these quantities are also possible via the manifold dependent stochastic Melnikov method, but Melnikov only applies to a very special class of models for which the unperturbed homoclinic orbit is available. Our methods are unique in that they can essentially be considered as a "black-box" of tools which can be applied to a wide range of stochastic dynamical systems in the absence of a priori knowledge of manifold structures. We use here a model of childhood diseases to showcase our methods. Our tools will allow us to make specific observations of: (1) loss of reducibility between basins with increasing noise, (2) identification in the phase space of active regions of stochastic transport, (3) stochastic flux which essentially completes the heteroclinic tangle.

Original languageEnglish
Pages (from-to)153-177
Number of pages25
JournalPhysica D: Nonlinear Phenomena
Volume173
Issue number3-4
DOIs
StatePublished - 15 Dec 2002

Fingerprint

dynamical systems
operators
characteristic equations
escape
boxes
chaos
orbits
approximation

Keywords

  • Epidemiology
  • Measles
  • Noise
  • Stochastic Frobenius-Perron operator
  • Stochastic bifurcation
  • Stochastic chaos
  • Stochastic dynamical systems
  • Stochastic heteroclinic tangle
  • Stochastic transport

Cite this

@article{7613c32034fc47a880626578c7b5842d,
title = "A manifold independent approach to understanding transport in stochastic dynamical systems",
abstract = "We develop a new collection of tools aimed at studying stochastically perturbed dynamical systems. Specifically, in the setting of bi-stability, that is a two-attractor system, it has previously been numerically observed that a small noise volume is sufficient to destroy would be zero-noise case barriers in the phase space (pseudo-barriers), thus creating a pre-heteroclinic tangency chaos-like behavior. The stochastic dynamical system has a corresponding Frobenius-Perron operator with a stochastic kernel, which describes how densities of initial conditions move under the noisy map. Thus in studying the action of the Frobenius-Perron operator, we learn about the transport of the map; we have employed a Galerkin-Ulam-like method to project the Frobenius-Perron operator onto a discrete basis set of characteristic functions to highlight this action localized in specified regions of the phase space. Graph theoretic methods allow us to re-order the resulting finite dimensional Markov operator approximation so as to highlight the regions of the original phase space which are particularly active pseudo-barriers of the stochastic dynamics. Our toolbox allows us to find: (1) regions of high activity of transport, (2) flux across pseudo-barriers, and also (3) expected time of escape from pseudo-basins. Some of these quantities are also possible via the manifold dependent stochastic Melnikov method, but Melnikov only applies to a very special class of models for which the unperturbed homoclinic orbit is available. Our methods are unique in that they can essentially be considered as a {"}black-box{"} of tools which can be applied to a wide range of stochastic dynamical systems in the absence of a priori knowledge of manifold structures. We use here a model of childhood diseases to showcase our methods. Our tools will allow us to make specific observations of: (1) loss of reducibility between basins with increasing noise, (2) identification in the phase space of active regions of stochastic transport, (3) stochastic flux which essentially completes the heteroclinic tangle.",
keywords = "Epidemiology, Measles, Noise, Stochastic Frobenius-Perron operator, Stochastic bifurcation, Stochastic chaos, Stochastic dynamical systems, Stochastic heteroclinic tangle, Stochastic transport",
author = "Bollt, {Erik M.} and Lora Billings and Schwartz, {Ira B.}",
year = "2002",
month = "12",
day = "15",
doi = "10.1016/S0167-2789(02)00659-0",
language = "English",
volume = "173",
pages = "153--177",
journal = "Physica D: Nonlinear Phenomena",
issn = "0167-2789",
publisher = "Elsevier",
number = "3-4",

}

A manifold independent approach to understanding transport in stochastic dynamical systems. / Bollt, Erik M.; Billings, Lora; Schwartz, Ira B.

In: Physica D: Nonlinear Phenomena, Vol. 173, No. 3-4, 15.12.2002, p. 153-177.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A manifold independent approach to understanding transport in stochastic dynamical systems

AU - Bollt, Erik M.

AU - Billings, Lora

AU - Schwartz, Ira B.

PY - 2002/12/15

Y1 - 2002/12/15

N2 - We develop a new collection of tools aimed at studying stochastically perturbed dynamical systems. Specifically, in the setting of bi-stability, that is a two-attractor system, it has previously been numerically observed that a small noise volume is sufficient to destroy would be zero-noise case barriers in the phase space (pseudo-barriers), thus creating a pre-heteroclinic tangency chaos-like behavior. The stochastic dynamical system has a corresponding Frobenius-Perron operator with a stochastic kernel, which describes how densities of initial conditions move under the noisy map. Thus in studying the action of the Frobenius-Perron operator, we learn about the transport of the map; we have employed a Galerkin-Ulam-like method to project the Frobenius-Perron operator onto a discrete basis set of characteristic functions to highlight this action localized in specified regions of the phase space. Graph theoretic methods allow us to re-order the resulting finite dimensional Markov operator approximation so as to highlight the regions of the original phase space which are particularly active pseudo-barriers of the stochastic dynamics. Our toolbox allows us to find: (1) regions of high activity of transport, (2) flux across pseudo-barriers, and also (3) expected time of escape from pseudo-basins. Some of these quantities are also possible via the manifold dependent stochastic Melnikov method, but Melnikov only applies to a very special class of models for which the unperturbed homoclinic orbit is available. Our methods are unique in that they can essentially be considered as a "black-box" of tools which can be applied to a wide range of stochastic dynamical systems in the absence of a priori knowledge of manifold structures. We use here a model of childhood diseases to showcase our methods. Our tools will allow us to make specific observations of: (1) loss of reducibility between basins with increasing noise, (2) identification in the phase space of active regions of stochastic transport, (3) stochastic flux which essentially completes the heteroclinic tangle.

AB - We develop a new collection of tools aimed at studying stochastically perturbed dynamical systems. Specifically, in the setting of bi-stability, that is a two-attractor system, it has previously been numerically observed that a small noise volume is sufficient to destroy would be zero-noise case barriers in the phase space (pseudo-barriers), thus creating a pre-heteroclinic tangency chaos-like behavior. The stochastic dynamical system has a corresponding Frobenius-Perron operator with a stochastic kernel, which describes how densities of initial conditions move under the noisy map. Thus in studying the action of the Frobenius-Perron operator, we learn about the transport of the map; we have employed a Galerkin-Ulam-like method to project the Frobenius-Perron operator onto a discrete basis set of characteristic functions to highlight this action localized in specified regions of the phase space. Graph theoretic methods allow us to re-order the resulting finite dimensional Markov operator approximation so as to highlight the regions of the original phase space which are particularly active pseudo-barriers of the stochastic dynamics. Our toolbox allows us to find: (1) regions of high activity of transport, (2) flux across pseudo-barriers, and also (3) expected time of escape from pseudo-basins. Some of these quantities are also possible via the manifold dependent stochastic Melnikov method, but Melnikov only applies to a very special class of models for which the unperturbed homoclinic orbit is available. Our methods are unique in that they can essentially be considered as a "black-box" of tools which can be applied to a wide range of stochastic dynamical systems in the absence of a priori knowledge of manifold structures. We use here a model of childhood diseases to showcase our methods. Our tools will allow us to make specific observations of: (1) loss of reducibility between basins with increasing noise, (2) identification in the phase space of active regions of stochastic transport, (3) stochastic flux which essentially completes the heteroclinic tangle.

KW - Epidemiology

KW - Measles

KW - Noise

KW - Stochastic Frobenius-Perron operator

KW - Stochastic bifurcation

KW - Stochastic chaos

KW - Stochastic dynamical systems

KW - Stochastic heteroclinic tangle

KW - Stochastic transport

UR - http://www.scopus.com/inward/record.url?scp=0037114536&partnerID=8YFLogxK

U2 - 10.1016/S0167-2789(02)00659-0

DO - 10.1016/S0167-2789(02)00659-0

M3 - Article

AN - SCOPUS:0037114536

VL - 173

SP - 153

EP - 177

JO - Physica D: Nonlinear Phenomena

JF - Physica D: Nonlinear Phenomena

SN - 0167-2789

IS - 3-4

ER -