Math 4/5027 - Nonlinear Dynamics and Chaos
Problem Set #4 - Spring 1998
Due March 23
Please do not postpone the assignment until the last minute!
Remember you also need to identify a Challenge problem (one for
undergraduates and two for graduate students) to work before the end of
the semester. * marks problems for graduate students (extra credit for
undergraduates).
-
Reading. By popular demand, we will move on to discuss continuous
problems (ODEs). We will cover all of Chapter 7; in fact, much of it
should be review, particularly for those who had the modelng course. I
will survey the chapter and point out the highlights in class. Please
read Chapter 7 in its entirety.
Please do the following problems.
- Preservation of containment. A key result in using
transition graphs is that maps preserve set containment. Prove that if
f is a continuous map on the real line, then
implies
that
. (Keep it short and use the definition of the
image of a set:
where
.) - Periodic points. Consider the graph of a map given at the
end of the asignment. Use a transition graph to determine the order of
all possible periodic points of the map.
- Period three and chaos. Consider the transition graph that
allows only the path
, and hence a
period three orbit. Does period three imply chaos in this case? - *You knew it was coming. Derive the conjugate map for the
tent map T and the logistic map
. First, try every approach
you can think of; perhaps you can devise a method different than the
one below.
- Consider the interval
and use the
definition of the conjugate map, G(C(x))=C(T(x)). Write the
functional equation for C(x) that results. - There is no standard approach to solving functional equations. One
can look for special forms of the solution that exploit apparent
properties of the solution (for example, polynomials, power series,
rational functions, exponentials, Fourier series). Notice that the
functional equation for C(x) involves
and C(2x), where
now means C squared, not C composed with itself. (If your
equation doesn't have these terms, return to part (a).) What common
family of functions has the property that
? - Given the insight from part (b), you might be tempted to look for
solutions of the form
, where
and
are to be determined. This looks like one term of a Fourier
series and you will see that it's easier to work with
than
sines and cosines. In order to make C(x) real-valued, it would be best
to use
, where
is the
complex conjugate of a. Show that when this trial solution is
substituted into the functional equation, no useful information can be
obtained about a and k. - Why would real exponentials,
, make matters even worse
than complex exponentials
? - The approach in part (c) almost worked and suggests that the trial
solution may need another term. We have included sines and cosines of
arbitrary frequency k, so a qualitatively different term is a
constant. Try a trial solution of the form
Now determine a and b. Can k be determined?
- We know that C must be one-to-one and map [0,1] onto [0,1],
so it is reasonable to ask that C(0)=0 and C(1)=1. Furthermore, to
impose some symmetry, let's ask that
. What
constraints do these conditions place on k? Find the most general form
of C(x).
- * Explicit formula for the logistic map. Recall that in
Chapter 1, problem 1.15, an explicit formula was given for the sequences
generated by the logistic map. Use the conjugate map to derive this
formula for the case that
. - Phase plane problems. Please do problem 7.2 (Chapter 7).
- Phase plane problems. Please do problem 7.5(c). (The sketch
can be fairly rough.)
- Predator-prey model. Please do problem 7.13(a).