Historical Overview ODE’s Def, Geometric approach 1st Examples Dimensional impact 1D
Dynamical Systems [1]
Jeandrew Brink, UFS
July 2021
Dynamical Systems [1] Jeandrew Brink, UFS
Historical Overview ODE’s Def, Geometric approach 1st Examples Dimensional impact 1D
Historical Overview
1950 Computer invented
1666 Newton - invented ODE’s to
explain planetary motion, 1963 Lorenz - chaotic system in
Analytic studies, mechanics model of atmosphere
(convection rolls)
1890 Poincare - 3-body problem
unsolvable, geometric 1970’s May chaos in pop biology,
approach, ”visions of chaos” iterated maps xn+1 = f (xn )
Mandelbrot - fractals
1920-50 Non-linear oscillators in physics (geometric chaos)
and engineering, radar, radio, Winfree - nonlinear oscillators
phase locked loops, lazer. in biology
1920-60 Birkhoff Kolmogorov Arnold Ruelle and Takens -
Moser, KAM theorem - chaos → turbulence
complex behaviour in Feigenbaum -universal route to
Hamiltonian dynamics. chaos, connection to statistical
physics, renormalization
Dynamical Systems [1] Jeandrew Brink, UFS
Historical Overview ODE’s Def, Geometric approach 1st Examples Dimensional impact 1D
Logical structure of Dynamics
Let x ≡ (x1 , · · · , xn ) be a
coordinate vector of n components, GOAL: understand how x1 , · · · xn
then the system of ordinary depend on time, t.
differential equations (ODE’s) in
Old strategy find formula for
autonomous first order form is
expressed in vector form as x1 (t), · · · , xn (t) given initial
ẋ = f (x) where ẋ ≡ dx conditions x1 (0) · · · xn (0)
dt .
Equivalently ... New approach - visualize
ẋ1 = f1 (x1 , · · · , xn ) (x1 (t) · · · xn (t)) geometrically.
.. ..
. = .
ẋn = fn (x1 , · · · , xn )
Where f1 , · · · fn are given functions.
Dynamical Systems [1] Jeandrew Brink, UFS
Historical Overview ODE’s Def, Geometric approach 1st Examples Dimensional impact 1D
Ex 1. Simple harmonic oscillator
mẍ + kx = 0
Write it in first order form
Let x1 = x and x2 = ẋ and then rewrite
the ODE ito x1 , x2
x˙1 = ẋ = x2 and x˙2 = ẍ
x˙ = x2
First order form: =⇒ 1 k
x˙2 = − m x1
Eq. is linear. (Only 1st power of
components on found in f ). The sin function on the RHS makes the
Linear systems can in general be
equation non-linear. n = 2, so 2D
rewritten as ẋ = Ax where A
is and n × n constant coefficient matrix. system. If the sol x1 (t), x2 (t) were
Can be solved in general ... known, we could draw the picture in
Ex 2. Simple Pendulum (x1 , x2 ) phase space as the point moves
Governing Eq. θ̈ + (g /l) sin θ = 0 along the trajectory. Seek the ”phase
To 1st order form x1 = θ, x2 = θ̇ portrait” - picture of all trajectories -
x˙1 = x2 , x˙2 = −(g /l) sin x1 without solving the eqs.
Dynamical Systems [1] Jeandrew Brink, UFS
Historical Overview ODE’s Def, Geometric approach 1st Examples Dimensional impact 1D
Impact of Dimensionality
Dim Linear Non-linear
n=1 Exponential growth / decay Fixed Points. Bifurcations.
n=2 Linear Ossilator Pendulum, vd Pol
2 Body prob (Kepler) Limit Cycles,
n≥3 Civil Structural Engineering Chaos. Lorenz, Poincare, ...
Electrical Engineering Turbulence, fire-flies Lazers ...
Dynamical Systems [1] Jeandrew Brink, UFS
Historical Overview ODE’s Def, Geometric approach 1st Examples Dimensional impact 1D
One dimensional systems (n = 1)
ẋ = f (x) where ẋ = dxdt , x ∈ R
Let x ≡ position of an imaginary
Ex. ẋ = sin x. particle moving along a line.
dx and ẋ ≡ velocity of that particle.
Solve analytically: Rsin x = dt Then ẋ = sin x acts as a rule for a
=⇒ t + Const = cosec(x)dx velocity of a particle when it is at x.
= − ln |cosec(x) + cot(x)|
1 ẋ
To evaluate the Const.:
suppose x = x0 at t = 0. x
cosec(x)+cot(x)
then t = − ln cosec π x0 π π 3π 2π 5π
(x0 )+cot(x0 ) 2 2 2 2
UGLY :-( !!! Hard to see what −1 ẋ = sin x
x is like as a function of t. If x0 = π4 , x(t) → π as t → ∞
If x0 = π4 , what is limt→∞ x(t)? Particle moves right when ẋ > 0 and
Much easier to use a geometric left when ẋ < 0 stationary when
approach. Gives greater insight. ẋ = 0 ≡ “fixed points” denote x ∗
Dynamical Systems [1] Jeandrew Brink, UFS