Solving First Order Ordinary Differential Equations Using Lie Symmetries

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 22

Solving First Order Ordinary

Differential Equations Using


Lie Symmetries

Katie Thompson
Overview
 Discrete symmetries vs. continuous
symmetries.
 What is a Lie symmetry?
 Using Lie symmetries to solve first order
ordinary differential equations.
 Examining “standard methods” taught
in a first course in ordinary differential
equations.
Transformations that are
Symmetries
 We look at diffeomorphisms on the
plane, i.e., maps :R2 R2.
 (x,y)=(f(x,y),g(x,y)), which are
invertible and infinitely differentiable.
 The set of diffeomorphisms {} forms a
group under composition.
Symmetries of Differential
Equations
 A symmetry of a differential equation is
a transformation that sends solutions to
solutions.
 The identity map is a trivial example of a
symmetry.
 Let y’(x)=(y). Then a translation in x
sends solutions to solutions.
Discrete vs. Continuous
Symmetries
y
 A rigid square has a
finite, discrete set of
symmetries.  x
 We interest
ourselves in infinite
sets of continuous
Any rotation by  R about the
symmetries. center is a symmetry. We denote
 Consider the unit the set of symmetries:
circle: x2+y2=1. {(x,y)}={(xcos-ysin,xsin+xcos)}.
One-Parameter Lie Group
Properties
 The set of diffeomorphisms {} for
 R forms a one-parameter Lie group if
the following hold:
  = 0 is the trivial symmetry.
 =+.
 Notice that the above imply that
-=-1.
Example of a One-Parameter
Lie Group
y

 Again, consider the  x


unit circle: x2+y2=1.

Any rotation by  R about the


center is a symmetry. We denote
the set of symmetries:
{(x,y)}={(xcos-ysin,xsin+xcos)}.
Symmetry Condition
 For the first-order ODE y’(x)=(x,y), a
symmetry group preserves the
equation, i.e.

 This is equivalent to the condition:


Example

Scaling is a symmetry of this ODE.

Verify the symmetry condition:


Goal
 We want to use the Lie symmetries to
separate variables and find the general
solution for a first-order ordinary
differential equation.
 From the symmetries, we find canonical
coordinates. The canonical coordinates
are the key to separating variables.
Infinitesimal Generators
 Let p=(x,y)be a point in R2 and let {}
be a one-parameter Lie group.
 Then the orbit of p under the group
action is the set, Op={(p)| R}.
 Then we define the infinitesimal
generator of  at the point p as the
tangent vector to Op at =0.
Canonical Coordinates
 We want to calculate canonical
coordinates, (r,s)=(r(x,y),s(x,y)).
 These canonical coordinates are built so
that the Lie symmetries become
translations (r,s+).
 Then the ODE reduces to
Canonical Coordinates
 We know that for the translation,
(r,s+), the tangent vector at (r,s) is
given by (0,1).
 We use the chain rule to get:
Canonical Coordinates
 We can find r by solving:

 Solve r(x,y) for y.


 We find s by calculating:

 Then evaluate the integral at r=r(x,y).


Solve ODEs Using Canonical
Coordinates
 To reduce the ODE to separation of
variables, rewrite in terms of canonical
coordinates.
First-Order ODEs of
Homogeneous Type
 These are differential equations of the
form:

 Make the substitution z(x)=y(x)/x.


 Use this substitution to solve the ODE.
 This method comes from the symmetry:
Canonical Coordinates
 Then canonical coordinates are given by

 We rewrite in terms of canonical


coordinates to get:
Example of 1st Order ODE of
Homogeneous Type
Riccati Equation
Solutions to Riccati Equations
 Apply the transformation (ye-2, xe) to
the solution when c is nonzero.

 When c=0, the transformation sends


the curve to itself.
Beyond 1st Order ODEs
 Higher Order ODEs: Symmetries are
used to reduce the order of the
differential equation.
 Partial Differential Equations:
 Methods to find symmetric solutions are
well understood.
 General methods to find non-symmetric
families of solutions are not known, and is
an open area of research.
References
 Edwards, C. Henry and Penney, David
E. Differential Equations: Computing
and Modeling. Upper Saddle River, NJ:
Prentice Hall 2000.
 Hydon, Peter E. Symmetry Methods for
Differential Equations: A Beginner’s
Guide. Cambridge, UK: Cambridge
University Press, 2000.

You might also like