Scribd
Upload
42 views9 pages

MA417 Lecture 24

IIT Bombay

Uploaded by

Jittu Yadav
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
42 views9 pages

MA417 Lecture 24

IIT Bombay

Uploaded by

Jittu Yadav
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

MA-417-ODE

Lecture 24

Debanjana Mitra

Department of Mathematics
Indian Institute of Technology Bombay
Powai, Mumbai - 76

October 12, 2021


Interpretation of solutions of ODEs
Preliminaries
We recall few notions below:
Path: A continuous map γ : [a, b] → Rm is called a path in Rm .
If a path γ is a C 1 -map, then the path is called a ‘smooth path’.
For a smooth path γ, a tangent vector to γ at t = t0 is defined by γ 0 (t0 ).
Curve: The image of a path in Rm is called a curve in Rm , i.e., if
γ : [a, b] → Rm is a continuous map denoted the path, then the curve is
given by {γ(t) | t ∈ [a, b]}.
Note: Two different paths can give same curve.
Ex. γ1 (t) = (cos(t), sin(t)), t ∈ [0, 2π] and
γ2 (t) = (cos(2t), sin(2t)), t ∈ [0, π2 ] two different paths. But both give
same curve
C : x12 + x22 = 1, ∀ (x1 , x2 ) ∈ R2 .

Graph of a continuous map: Let φ : I → Rn be a continuous function,


where I ⊂ R. The graph of φ is defined by the set {(t, φ(t)) | t ∈ I }.
The map γ(t) = (t, φ(t)) for all t ∈ I is the associated path in R × Rn .
Goal to study the qualitative behavior of solutions of

Y 0 (t) = f (t, Y (t)),

where f : I × Ω → Ω, and I ⊂ R and Ω is an open subset of Rn .


Recall a solution of the above problem means that there exist an interval
I0 and a differentiable functions φ : I → Rn such that
1. (t, φ(t)) ∈ I × Ω for all t ∈ I0 ,
2. φ0 (t) = f (t, φ(t)), ∀ t ∈ I0 .
I The graph of φ in R × Rn , i.e., {(t, φ(t)) | t ∈ I0 } is called the
integral curve.
I To obtain a solution φ of the ODE means that there exists a
differentiable map γ associated to the graph of φ i.e, γ(t) = (t, φ(t))
for all t ∈ I0 such that γ 0 (t) = (1, f (γ(t)) for all t ∈ I0 .
I The set Ω in Rn is called the phase space. For n = 2, it is called
phase plane.
I The projection of the graph onto Ω, i.e., {φ(t) | t ∈ I0 } (i.e.,
Image(φ)) is called a trajectory or phase curve.
I Collection of all phase curves in phase space is called phase portrait.
Autonomous system
Let Ω be an open subset of Rn and let f : Ω → Rn be a given function.
Consider the system of ODEs:
Y 0 (t) = f (Y (t)), t ∈ R.
In the above system f does not depend on t and in fact f : Ω → Rn ,
where Ω ⊂ Rn . This system is called Autonomous system. Otherwise,
the system is called non-autonomous system.
Ex.
I Autonomous system: Y 0 (t) = AY (t) = f (Y (t)), where f (X ) = AX
for any X ∈ Rn and An×n is a matrix with constant entities.
I Autonomous system: In n = 1, consider y 0 (t) = y 2 (t) = f (y (t)),
where f (x) = x 2 for any x ∈ R.
I Non-autonomous system: In n = 1, for f (t, x) = t 2 x 2 for any t ∈ R
and x ∈ R, consider y 0 (t) = t 2 y 2 (t) = f (t, y (t)),
 3tx − 2x 
I Non-autonomous system: For f (x1 , x2 ) = 1 2
.
5x1 + t 2 x2

y10 (t) = 3ty1 (t) − 2y2 (t)


y20 (t) = 5y1 (t) + t 2 y2 (t).
Autonomous system: Let Ω be an open subset of Rn and let f : Ω → Rn
be a given function. Consider the system of ODEs:

Y 0 (t) = f (Y (t)), t ∈ R. (1)

Lemma
Let φ(·) be a solution of the autonomous system (1) for t ∈ (a, b). For
any c ∈ R, set ψ(t) = φ(t + c) for all t ∈ (a − c, b − c). Show that ψ(·)
is a solution of (1) on (a − c, b − c).
Proof. Using direct calculation,
ψ 0 (t) = φ0 (t + c) = f (φ(t + c)) = f (ψ(t)) for all t ∈ (a − c, b − c).

The above result is not true for non-autonomous system.


Ex. Check the above lemma does not hold for ths system: y10 (t) = y1 (t),
y20 (t) = ty1 (t).
Let us assume that f : Ω → Rn is a C 1 function and for all X ∈ Ω, the
system of ODEs: Y 0 (t) = f (Y (t)), Y (0) = X0 admits a unique
solution φ(t) for all t ∈ R. Show that for any t0 ∈ R, there exists a
unique solution ψ defined on R to the problem

Y 0 (t) = f (Y (t)), Y (t0 ) = X0 .

In fact, ψ can be obtained by ψ(t) = φ(t + t0 ) for all t ∈ R. (Check).

Remark: Thus without loss of generality for autonomous system, it is


enough to study the system of ODEs: Y 0 (t) = f (Y (t)), Y (0) = X0 ,
for all t ∈ R, where f : Ω → Rn is a C 1 function and the ODE has a
unique global solution.
Dynamical system
Smooth Dynamical system:A C 1 -map φ : R × Ω → Ω is said to be a
smooth Dynamical system if for each t ∈ R, the restricted function
φt : Ω → Ω defined by φt (X ) = φ(t; X ) satisfies the following properties:
1. φ0 : Ω → Ω is an identity map i.e., φ0 (X ) = X , ∀ X ∈ Ω.
2. (φt ◦ φs )(X ) = φt+s (X ), ∀ X ∈ Ω, ∀ t, s ∈ R.
Observe: taking s = −t in 2. above, we get (φt ◦ φ−t )(X ) = X and then
(φ−t ◦ φt )(X ) = X . Thus, φt is an invertible map whose inverse is φ−t , a
C 1 -map.
Ex. Φ(t; X ) = e tA X , where An×n is a matrix with constant entities.
Show that Φ(t; X ) is a smooth dynamical system.

Remark
I If φ is a dynamical system, then φ satisfies a system of ODE:
Y 0 (t) = f (Y (t)), where f : Ω → Rn defined by
d
f (X ) = dt φ(t, X )|t=0 .
I Conversely, a solution of a system of ODEs: Y 0 (t) = f (Y (t)) for a
given f : Ω → Rn , C 1 -function, is a dynamical system on Ω if and
only if for all X ∈ Ω, IVP with initial condition Y (0) = X admits a
unique solution φ(t; X ) for all t ∈ R.
Flow of a vector field

Let us assume that f : Ω → Rn is a C 1 function and for all X ∈ Ω, the


system of ODEs: Y 0 (t) = f (Y (t)), Y (0) = X admits a unique solution
φ(t; X ) for all t ∈ R.
Recall: For a fixed X ∈ Ω, the image of the map t 7→ φ(t; X ) in Ω is
called the trajectory.
Definition: For a fixed t ∈ R, the image of the map X 7→ φ(t; X ) in Ω
from Ω is called the flow of the dynamical system at time t.

In particular, the flow of a dynamics is the set


{φ(t; X ) | t ∈ R, X ∈ Ω}.
Recall that the function (i.e., X 7→ f (X )) defines a vector field in Ω. The
flow defined above is called the flow of the vector field f .

You might also like