Ordinary Differential Equations.

Robert Devany, Boston University (Chair)

Robert Borelli, Harvey Mudd College
Martha Abell, Georgia Southern University
Talitha Washington, Howard University

Introduction. This is a report from the working group charged with making
recommendations for the undergraduate curriculum in differential equations.
As discussed below, the basic sophomore level Differential Equations course
has changed dramatically (at least at some institutions) over the past twenty
years, mainly due to the much wider availiblility of computer resources. We
discuss a number of these changes below and then provide three different
syllabi, each aimed at a slightly different audience. Finally, we provide some
references including sample texts.

How Differential Equations Courses Have Evolved. As part of our

committee work, we surveyed a half dozen different schools regarding their
sophomore-level Ordinary Differential Equations (ODE) course. These schools
included two-year colleges, liberal arts colleges, and universities. Two things
became clear from this survey. The first is that the ODE course has under-
gone a remarkable transformation over the past twenty years. The second is
that there is now no set curriculum for this course.

Regarding the first point, the ODE course previously focused almost exclu-
sively on specific analytic methods for solving differential equations. Often,
examples of differential equations were simplified so that the correspond-
ing equation could then be solved explicitly (as, for example, the nonlinear
pendulum equation was often linearized to accomplish this). Usually, the
types of equations covered in the course were first and second (and maybe
higher) order differential equations but rarely included systems of differential
equations.

The major change in the ODE courses over the past two decades has been
primarily motivated by the availability of the computer. Currently, most

1
courses use some of the many different software packages that are available
to approximate and visualize solutions of ODEs. Differential equations that
are completely unsolvable using analytic techniques can now be approached
from a geometric point of view using tools from dynamical systems. These
include important differential equations that arise in many different areas of
science and engineering, topics that were untreatable in this course a few
years back. This has led to many new topics being introduced into this
course. As a consequence, some specialized analytical techniques for solving
ODEs have necessarily been dropped from the course.

The ODE Audience. At many institutions, the ODE course is often the
final mathematics course required by students in engineering and certain
areas of science. At many large universities, over 90% of students enrolled
in this course come from these client disciplines. Indeed, all of calculus (and
many topics from earlier mathematics courses) come together in this course,
and this helps demonstrate how important many of these earlier topics are
for students to learn. Of course, there are variants of the ODE course that
are aimed at a more mathematically oriented group of students; we include
such a syllabus below.

The Computational Approach. This new approach to teaching and learn-

ing ODEs has a number of advantages associated with it. A major advantage
of the computational approach is that modeling/research projects involving
real-life nonlinear systems of ODEs that arise in various areas of science and
engineering can now be included as part of the course. As mentioned above,
the ODE course often serves as the capstone mathematics course for students
in areas of science and engineering, although, of course, most mathematics
students go well beyond this course. Given that this is the final mathemat-
ics course for many students, it is natural that such students complete their
mathematical training by becoming involved in modeling/research topics that
are related to their specific major. Almost all such projects involve some sort
of numerical experimentation with ODEs that were previously inaccessible
to students. Modeling/research projects also develop the important skill of
visualizing solutions for ODE students.

These projects can be a team-effort or an individual assignment, and they

usually require an extensive written report. Time permitting, oral reports

2
on the project can also be required. Fortunately, almost all ODE textbooks
nowadays contain topics for such projects, and there is also plenty of infor-
mation on modeling applications and lab projects on-line. For example, see
the CODEE website.

New Topics to Include. Another advantage of the computational ap-

proach is that this allows for the inclusion of topics that have not usually
been treated in an introductory ODE course. For example, since visualiza-
tion of solutions is much more important these days, it is natural to convert
second order linear differential equations to planar systems of differential
equations, so solutions can then be visualized in the phase plane. Con-
sequently, linear systems of the form Y 0 = AY are now included in most
courses. Usually, only planar systems are covered, so A is a 2 by 2 matrix,
although more advanced courses often introduce higher dimensional systems
as well. Therefore the new course includes a relatively brief introduction to
linear algebraic topics such as matrices, eigenvalues, and eigenvectors. Often
these topics are introduced gradually during the course, not necessarily as a
week or two week long diversion into linear algebra. The inclusion of these
topics connects well with the linear algebra course that many students often
take after the ODE course, as it provides a nice visual application of linear
algebra. For example, the most important solutions of linear systems are the
“straight line” solutions. These are solutions of the form eλt V , where λ is an
eigenvalue of A and V is its associated eigenvector, so students see early on
why eigenvalues and eigenvectors areimportant.

In addition, nonlinear systems are often covered and this necessitates using a
qualitative approach to understanding the system as most nonlinear systems
cannot be solved analytically. Topics such as nullclines and linearization near
equilibria provide some of the tools necessary for this geometric approach.
Hamiltonian and gradient systems may also be introduced, and this provides
further tools. Some courses even occasionally include modern topics such as
chaos and bifurcation theory.

Computational Tools. There is no set standard for how to use computers

in the ODE courses. Many institutions use computer algebra systems such as
Maple or Mathematica. This is fine when students already have some famil-
iarity with these tools, but often certain students in the course do not come

3
with this background and hence they encounter a steeper learning curve.
Another resource is that many of the newer ODE texts come with software
specifically designed for topics included in the book. The advantage here
is that students do not have to take time to learn how to use the software.
A third possibility is the use of spreadsheets. When students try to learn
the numerical algorithms for approximating solutions to ODEs (like Euler’s
Method or Runge-Kutta 4), a natural method to encode these algorithms
involves a spreadsheet. Moreover, the graphical capabilities of spreadsheets
also help students visualize the outputs of these algorithms. A major advan-
tage of this approach is that almost all students in the ODE course already
have a good background using spreadsheets.

What to Eliminate. Given the inclusion mentioned above of more mod-

eling/research projects, topics from linear algebra, and more computational
based topics in the contemporary ODE courses, clearly some topics from the
traditional ODE course must be dropped. Most modern courses now elim-
inate some or many of the specialized analytic methods for solving ODEs.
For example, integrating factors, variation of parameters, solutions of special
equations like the Bernoulli equation, and other such techniques are some-
times not included. Instead, in some instances, the only types of first order
ODEs that are now solved analytically are linear and separable equations.
Also, series solutions are sometimes eliminated (though, of course, the lin-
earization techniques mentioned above do involve the “first” terms in such
a series solution). And, except in courses heavily populated by engineering
students, Laplace transforms can also be eliminated.

Cognitive Learning Goals

A. The Culmination of Calculus. ODEs really form the primary basis for
the study of calculus and so this course should strive to bring together
many of the previously covered concepts in a way that confirms their
usefulness and necessity.

B. Use of Technology. The ODE course is easily the course in the in-
troductory undergraduate mathematics curriculum in which the use of
technology is most essential. Students should be encouraged to use

4
 these tools in homework, in projects, and in simply visualizing the
various qualitative aspects of ODEs.

C. Applications. There are major applications involving differential equa-

tions in all areas of science and engineering, and so many of these should
be included in the ODE course to show students the relevance and im-
portance of this topic. Some applications include mass-spring systems,
forced, damped, and undamped pendulum equations, and Newton’s
Laws (physics), electrical circuits (engineering), enzymatic reactions
(chemistry), population models (biology), Kepler’s Laws (astronomy),
compound interest models (economics), and the Lorenz system (mete-
orology).

D. Introduction to Higher Level Mathematics. This course also pro-

vides an opportunity for students to get a glimpse of some topics in
higher level mathematics courses. Examples include linear algebra
(solving linear systems of ODEs and linearization), numerical analy-
sis (understanding numerical methods such as Euler’s method or RK4
for approximating solutions of ODEs), real analysis (the existence and
uniqueness theorem), and dynamical systems (bifurcation theory and
chaos).

Sample Syllabi. Because of the variety of aforementioned topics and styles,

there is now no one set curriculum for an ODE course. Each of the schools we
surveyed had different approaches to the course. Below is a summary of the
syllabi of the three types of ODE courses offered at the collegiate level: the
captsone college course, the sophomore level service course, and the course
aimed primarily at math majors.

1. The capstone ODE course not necessarily aimed at partner disciplines.

Many, if not all, community colleges offer Differential Equations as
the capstone course in the mathematics department. Some time is
devoted to reviewing relevant topics in single variable calculus. The
course would not qualify as a replacement for a junior-level introductory
ODE course offered at a four-year college or a university. It might,
however, qualify as a replacement for a service ODE course offered

5
 in the freshman or sophomore years at those institutions. Here is an
outline of a syllabus for such a 13-week semester ODE course.

• First order ODEs: linear equations, radioactive decay model, sep-

arable equations, review of relevant topics from single-variable cal-
culus, slope fields, classifying equilibrium points of autonomous
first order ODEs, existence and uniqueness theorems, introduc-
tion to the use of solvers, Euler’s Method, population models. (4
weeks)
• Second order ODEs: solution techniques for linear ODEs with
constant coefficients, review of complex numbers, direction fields,
mass-spring model, forced and damped harmonic motion, beats
and resonance. (4 weeks)
• Planar systems: solution techniques for planar linear systems with
constant coefficients, an introduction to relevant topics from lin-
ear algebra (matrices, eigenvalues, eigenvectors), nonlinear au-
tonomous systems, classifying equilibrium points, phase portraits,
various models. (5 weeks)

2. An ODE course which is part of the math core required of science

and engineering students taken in the freshman and sophomore years.
That is, this course is really a service course. Unlike the previous
syllabus, this course will include many more applications in the client
disciplines. As in the earlier course, students will have not taken a
course on Linear Algebra prior to this course, so the syllabus allots
time to cover the linear algebra topics needed for solving linear systems.
Laplace transforms are included. The modeling projects emphasize the
usability of ODEs in other disciplines such as biology (predator-prey
and competing species models), electrical engineering (circuit theory
models), and physics (mechanical systems, the n-body problem). The
syllabus for such a 13-week semester course might look something like
this:

• First-order ODEs: Slope fields, separable equations, linear equa-

tions, nonlinear ODEs, existence and uniqueness, numerical solu-
tions, qualitative analysis of solutions. (3 weeks)

6
 • Second-order ODEs: Linear ODEs with constant coefficients, forced
and damped harmonic oscillator, beats and resonance, phase plane,
applications to springs, electrical circuits. (3 weeks)
• Laplace Transforms: Definition, solving initial value problems for
linear constant-coefficient second-order ODEs. (2 weeks)
• Linear Systems of ODEs: Solution of constant-coefficient undriven
linear systems using techniques of linear algebra, phase plane anal-
ysis for linear systems, solution of driven linear constant-coefficient
systems. (3 weeks)
• Nonlinear Systems of ODEs: Autonomous systems, planar sys-
tems, direction fields, stability of linear and non-linear autonomous
systems. (2 weeks)

3. The ODE course as a stand-alone course in the mathematics depart-

ment taken in the junior year (or perhaps as an honors-level course in
place of the aforementioned courses). The syllabus includes all top-
ics in (2.) with the exception of Laplace transforms, and more theory
is included throughout the course. Additional topics may include⊗ :
theP oincaré−BendixsonT heorem, nonlinearplanarautonomoussystems, limitcycles, stabili

Possible Textbooks:

Remark: The presence of a text on this list is not meant to imply

an endorsement of that text, nor is the absence of a particular
text from the list meant to be an anti-endorsement. The texts are
chosen to illustrate the sorts of texts that support various types
of courses. Please note that some of the books listed below were
written by the authors of this report.

1. Abell, M.L., Braselton, J.P., Introductory Differential Equations, Else-

vier, Inc.

2. Blanchard, P., Devaney, R. L., Hall, G. R. Differential Equations.

Fourth Edition. Cengage Learning.

7
 3. Borelli, R.L., Coleman, C.S., Differential Equations: A Modeling Per-
spective, John Wiley and Sons, Inc.

4. Boyce, W.E., DiPrima, R.C., Elementary Differential Equations, John

Wiley and Sons, Inc.

5. Braun, M. Differential Equations and Their Applications. Springer.

6. Edwards, H.C., Penney, D.E., Differential Equations: Computing and

Modeling, Pearson Publishing

7. Edwards, H.C., Penney, D.E., Elementary Differential Equations with

Boundary Value Problems, Pearson Publishing

8. Nagle, R.K., Saff, E.B., Snider, A.D., Fundamentals of Differential

Equations, Pearson Publishing

9. Zill, D.G., A First Course in Differential Equations with Modeling Ap-

plications, Brooks/Cole, Cengage Learning

More Advanced Textbooks:

10. Hirsch, M. W., Smale, S., Devaney, R. L. Differential Equations, Dy-

namical Systems, and an Introduction to Chaos. Third Edition. Aca-
demic Press.

11. Strogatz, S. Nonlinear Dynamics and Chaos. Westview Press.

Online Resources:

12. The CODEE Project

13. The Boston University ODE Project

14. ODE Software for MATLAB

8
 15. SIMIODE - Systemic Initiative for Modeling Investigations and Oppor-
tunities with Differential Equations

Articles from Mathematics Education Research:

16. Rasmussen, Chris and Oh Nam Kwon, An inquiry-oriented approach

to undergraduate mathematics, Journal of Mathematical Behavior26
(2007), 189-194.

This article appeared in a special issue of the Journal of

Mathematical Behavior. In addition to providing an overview
of the five articles in the issue, the authors highlight the the-
oretical background for an innovative approach in differential
equations called the Inquiry Oriented Differential Equations
(IO-DE) project. and provide a summary of two quantitative
studies done to assess the effectiveness of the IO-DE project
on student learning.