Professor Henry Greenside
Welcome
Prerequisites
Time and Place
Office Hours
Computer Accounts
Grading
References
Welcome
Welcome to
Physics 513, one of the core courses of Duke
University's
Center for Nonlinear and Complex Systems. This
course is an introduction to and survey of nonlinear
dynamics, an exciting and interdisciplinary area of
research concerning how natural and man-made systems
evolve in time. Besides being interesting in its own
right, nonlinear dynamics has proved useful to many
disciplines including biology, chemistry, computer
science, engineering, environmental science, geology,
mathematics, medicine, meteorology, physiology, and
physics.
A goal of the course will be to study the successes,
limitations, and implications of a modern discovery,
that simple deterministic nonlinear evolution
equations can generate complex behaviors that
quantitatively agree with experimental
observations of many physical systems. It will take a
while to give you the background to appreciate the
scope of this discovery and so the course will
introduce and discuss many of the following topics:
-
Concepts related to a geometric and global way of
thinking about nonlinear evolution equations. These
concepts include: phase space, dissipative versus
conservative systems, attractors, basins of attraction,
elementary bifurcation theory, linear stability theory,
Poincare sections and maps, strange attractors,
Lyapunov exponents, transition scenarios (Feigenbaum,
Ruelle-Takens, and intermittency), universality,
synchronization, fractals, fractal dimensions, and
analysis of time series by embedding.
-
Comparisons of theory with high-precision laboratory
experiments, e.g., the sequences of transitions that
lead to chaos in a convecting fluid of moderate lateral
extent.
- Applications of nonlinear dynamics to different
disciplines, e.g., ecology, engineering, neurobiology,
and fluid dynamics.
-
Strategies and algorithms for simulating, analyzing,
and controlling nonlinear systems, including the
integration of differential equations, and the
calculation of power spectra, Lyapunov exponents, and
fractal dimensions. We will also discuss fundamental
limitations imposed on numerical simulation by
nonlinear dynamics, e.g., the difficulties of accurate
long-term forecasting in the presence of chaos.
Students enrolled in 513 should attend the
biweekly
seminars
of the Center for Nonlinear and Complex Systems.
These seminars are generally held on Tuesdays in
Physics 113 from 2:15-3:15 pm. Students in
this course should also consider enrolling in the
course NCS 201/Physics 201.This course meets once
a week on Thursdays and is taught by a team of 3-4 CNCS
faculty, who in turn discuss their research and its
relation to nonlinear dynamics.
Prerequisites
The course is intended for advanced undergraduates and
graduate students with diverse backgrounds. Students
will need to have a working knowledge of multivariate
calculus and of linear algebra at a sophomore
undergraduate level. They should feel comfortable with
the concepts of limits, complex numbers (you should
especially be familiar with the identity exp(ix)
= cos(x) + i sin(x)), integrals,
derivatives, the gradient and divergence operators,
multivariate Taylor series, Fourier series, matrices,
determinants, eigenvalues and eigenvectors, big-oh
notation, and ordinary differential equations.
Students should have had at least one introductory
undergraduate physics course so that concepts such as
vectors, forces, momentum, energy, energy conservation,
flux, Newton's equations of motion, temperature, and
heat are familiar.
Some experience with a programming language like
Fortran, Java, or C++ is also needed although not
at a sophisticated level, e.g., the computer statement
"i = i + 1" should make sense to you. Some homework
assignments and lecture demonstrations will use
previously written programs for Mathematica, a
powerful computer program that provides an interactive
symbolic, numerical, and graphical environment for
mathematics and data analysis.
Although many of these mathematical, physical, and
computational concepts will be reviewed in class as
they are used, the review will usually be terse.
Students not comfortable with these concepts should
talk to me before enrolling.
Time and Place
The class will meet Monday, Wednesday, and Friday from
11:50 am-12:40 pm in Physics 158. There
will be occasional supplementary meetings, e.g., to
discuss homework assignments, to see and discuss a
video, or to tour the lab of some CNCS researcher.
At least an hour before each class, please get in the
habit of viewing the
Announcements section of the 513 home page. This
section will be updated frequently to mention possible
last minute changes in the class schedule, availability
of homework assignments and solutions, seminars and
colloquia of interest to the class, and the time and
place of supplementary classes.
Office Hours
I do not have fixed office hours. However, I will make
my best effort to meet with you if you have any
questions at all about the course (or more generally
about nonlinear dynamics or about physics or about Duke
of if you would just like to chat). If you are in the
Physics building, please free to drop by my office
Physics 047 and say hi.
To meet with me at some specific time, please email me
at the address
hsg@phy.duke.edu or call me at my Physics office
at 660-2548.
Feel free to send me e-mail at any time. I am often logged
on in the evenings and on the weekends and will be glad to
discuss the course or homework with you.
Computer Accounts
Students taking the course will need access to
electronic mail, to an Internet browser capable of
running Java applets, and to the computer mathematics
program Mathematica. (The latter is available to
students for free at this
Duke OIT link.) Lectures, homework assignments,
electronic documents, data sets, and graphical files
will be available through the
Miscellaneous Files link from the 513 home
page.
Grading
Your final grade for the course will be based on your
class participation, homework assignments, a midterm
exam, and a final project. These will be weighted
approximately as follows:
| Activity |
Percent of Total Grade |
| Class participation |
5% |
| Quizzes |
10% |
| Midterm exam |
20% |
| Final project (oral presentation and paper) |
25% |
| Homework assignments |
40% |
There will not be a final examination.
Class participation.
I expect all members of the class to read and to think
about the assigned material before lecture and to come
prepared to ask questions and to discuss the material
in class. If you don't understand something during
lecture or from the assigned reading, please don't be
shy, ask questions! If something catches your interest
and you want to learn more, ask questions. Talking with
me outside of lecture is also one way to participate in
class. I want to see evidence that you are actively
trying to learn and think about the course material.
Quizzes
There will be a 15-minute closed-book quiz roughly
every two weeks. The quizzes will consist of a few
short questions that you should be able to answer
easily if you are keeping up the material, so the
quizzes should give some useful feedback to you and to
me as to whether the lectures are being successful.
Midterm exam.
This will be an overnight 3-hour open-book exam that
will be given about half-way through the semester.
Final project of a presentation and paper.
Instead of a final exam, each member of the class will
write a 10-page double-spaced paper (including figures
and references) on some topic related to nonlinear
dynamics that he or she is especially interested
in. This paper can be based on reading some journal
article, on carrying out some simple experiment, or by
simulating or analyzing some mathematical model. Each
student will give a 15-minute oral presentation to the
class about their paper.
Homework assignments
There will be a homework assignment roughly every one
to two weeks, for a total of about seven
assignments. These are the heart of the course in that
it is through these assignments that you will build up
and apply your understanding of the various concepts
and techniques. Please keep the following in mind as
you work on and write up your assignments:
-
Late homework is not accepted unless you give me
a reasonable excuse at least three days before the
homework is due.
-
Your main two goals in writing up your homework are
to be clear (so that I can understand what you
have written) and to demonstrate
insight. Writing clearly means using readable
handwriting . You should avoid tiny script and avoid
trying to cram many sentences and equations onto a
single page. Leave plenty of space between symbols and
between successive lines of equations. Leave plenty of
space between the ending of one homework problem and
the beginning of the next. Spread your answers out over
many pages if necessary. (Paper is cheap compared to
the time needed for you to complete the assignments and
for me to grade your assignments.) If I can not read
and understand your assignments easily, you will get
little or no credit.
Demonstrating insight means using complete
sentences that explain what you are doing and
why. Cryptic brief answers like "yes", "no",
"24", or "f(x)" will not be given
credit. Instead, explain what you are doing and why,
e.g., as if to a friend who is not familiar with this
course. Your homework must show that you understand how
you got your answer and that you appreciate the
significance of your answer. A well-written complete
answer is one that you will be able to understand
yourself a month after you have written the answer,
even if you don't remember the original question.
-
You are allowed to collaborate on the homework
assignments (this is realistic, scientists collaborate
all the time in research) but as much as possible you
should attempt the assignments on your own since you
will learn the most that way. Whether or not you
collaborate, you must write up your homework on
your own, in your own words, and with your own
understanding. You must also acknowledge explicitly at
the beginning of your homework anyone who gave you
substantial help, e.g., classmates, myself, or other
people. (Again, scientists usually acknowledge in
their published articles colleagues that helped to
carry out the research.) Failure to write your
homeworks in your own words and failure to acknowledge
help when given can lead to severe academic penalties
so please play by the rules.
- The assignments will require typically a mixture
of analytical, numerical, and graphical approaches. The
mathematical derivations or analyses for the analytical
problems should be written out by hand on paper.
Please use ink, not pencil. Numerical and graphical
answers involve output that are best printed out on a
laserprinter, then stapled to your handwritten
sheets. A hand-sketch of a graphical plot with
essential features described is also acceptable.
-
Please pay attention to details as you write your
assignments. All symbols should be given names the
first time you introduce them, e.g., say "the momentum
p" or "the flux F" instead of just
using the symbols p and F. Physical units
should be given for any answer that is a physical
quantity, e.g., say "the angular momentum was
A=0.02 J-sec" or "the angle was
µ=0.32 radians." Numerical answers should
have the minimum number of significant digits that is
consistent with the given data. For example, if you
have a product or ratio of numbers of which the least
accurate number has two significant digits, the final
answer should have only two significant digits. Graphs
should have their axes clearly labeled by the
corresponding variables and by the variables' physical
units. Each graph should have a title that explains the
graph's purpose. A good way to learn how to write
effectively is to imitate the style of published
articles, e.g., those published in Physical Review Letters
.
-
If you use using Mathematica in a homework assignment,
please do not give me the output of your entire
session. Instead, just give me enough output to
convince me that you have answered the question
correctly. You should also include any Mathematica code
that you write so that I can try to understand how you
obtained your answers.
References
The course will be taught from my own notes and will
not follow any textbook particularly closely. One
required reference is the book Nonlinear Dynamics
and Chaos by Steven Strogatz (Addison-Wesley,
Reading, MA, 1994). This book is particularly well
written and accessible and should be the first place to
look for discussions and examples that complement my
lectures. However, this book is a bit too elementary
for the intended level of this course and does not
cover some of the topics in enough detail.
The following are some other books that I encourage you
to take advantage of during the semester and after. The
first three are on reserve in the Vesic Library
.
-
Chaos by James Gleick (Penguin, New York,
1987). A popular, entertaining and non-mathematical
survey of key advances in nonlinear dynamics and about
the people who made these advances. If you have no
prior knowledge about nonlinear dynamics, you should
read this book as soon as possible.
-
Chaos: An Introduction to Dynamical Systems by
K. Alligood, T. Sauer,and J. Yorke (Springer-Verlag,
New York, 1997). A book similar in level and content to
Strogatz but goes more deeply and carefully into
mathematical definitions and the derivation of results.
-
Chaos in Dynamical Systems, 2nd Edition by
Edward Ott (Cambridge University Press, 1993). This is
an advanced graduate-level discussion of chaos theory,
with many explicit mathematical examples worked out
with impressive insight.
-
Order within Chaos by Pierre Berge, Yves Pomeau,
Christian Vidal (John Wiley & Sons, New York, 1984). An
especially good introductory book for scientists (as
opposed to mathematicians) with many comparisons of
ideas with experimental data, but unfortunately
somewhat out of date and lacks exercises. The chapters
about the Fourier analysis of time series,
intermittency, and the synchronization of oscillators
are particularly good.
-
Introduction to Chaos by Michael Cross is an online
graduate-level course that is also especially well
suited for scientists and close in level and emphasis
to my course. Many good insights here, strongly
recommended. Also useful are many
Java demos that Mike wrote to complement his
lectures.
-
Nonlinear time series analysis by Holger Kantz
and Thomas Schreiber (Cambridge University Press,
1997). A good summary of concepts and techniques for
deducing information about a dynamical system from
empirical time series.
- Practical Numerical Algorithms for Chaotic
Systems by T. S. Parker and
L. O. Chua (Springer-Verlag, 1989). This book
discusses numerical methods for analyzing nonlinear
dynamical systems, although primarily lower-dimensional
ones.
Nonlinear dynamics remains an active and exciting
research frontier for many disciplines. You should get
in the habit of browsing weekly through key scientific
journals to get a sense of ongoing research. The
following journals are often quoted for research
related to nonlinear dynamics:
Top of this page
513 Home Page
Department of Physics