First Integrals Flashcards
Newton’s Equation
m x’’ = F(x)
-where the force F(x) depends only on the position of the point, but does not depend on time t or velocity v=dx/dt
First Integral
Definition
-a non-constant differentiable function |Ф = |Ф(|x,t) is called a first integral of the system: d|x/dt = |F(|x) , |x, |F(|x)ϵℝ^n IF dФ/dt = ∂Ф/∂t + Σ dxn/dt ∂Ф/∂xn = ∂Ф/∂t + |F(|x) . ∇Ф = 0
Time Independent First Integral
Definition
-we say that a first integral is time independent if:
∂Ф/∂t = 0
Geometric Meaning of First Integrals
-time independent first integrals Ф(|x) have simple geometrical interpretation, a trajectory passing through the point |xo stays on the surface:
V = {|xϵℝ^n | Ф(x)=Ф(xo)}
-which is a level set of the first integral
Level Sets and Integral Curves
- a level set of time dependent first integrals gives a surface in the extended phase space
- if one point of an integral curve belongs to this surface, then the whole integral curve belongs to it
What is the dimension of the level sets of a first integral for a system of N equations?
- for a system of N equations, the elements of the level set of a first integral are N-1 dimensional surfaces
- a trajectory of the system belongs to a level set
Where are the trajectories in systems with more than one first integral?
- a trajectory must simultaneously belong to the surface levels of each first integral
- i.e. it is a line of the intersection of the surfaces
First Integrals and Differentiable Functions
-if Ф=Ф(|x,t) is a first integral, then any differentiable function f(Ф) is a first integral since:
df(Ф)/dt = dФ/dt * df(Ф)/dФ = 0 since dФ/dt=0 by definition
-if a system has two or more first integrals then any differentiable function of these integrals is a first integral;
df(Ф1,Ф2,..,Фn)/dt
= dФ1/dt df/dФ1 + … + dФn/dt df/dФn = 0
Counting First Integrals
-counting first integrals we should count only functionally independent first integrals
Functionally Independent
Definition
-if Ф = f(Ф1,…,Фn) then the gradients of Ф and Ф1,…,Фn are linearly dependent:
∇Ф = ∇ f(Ф1,…,Фn)
= df/dФ1 ∇Ф1 + … + df/dФn ∇Фn
-note that the coefficients can be functions of dynamical variables x1,…,xn
-in other words, if gradients ∇Ф1, … , ∇Фn are linearly independent, then the first integrals Ф1,…,Фn are functionally independent
How to show that first integrals Ф1,…,Фm are functionally indepenent
1) compute their gradients
2) form an NxM matrix:
J(Ф1,..,Фm) = first row ∇Ф1, second row ∇Ф2, … , mth row ∇Фm
3) reduce to REF to determine the rank of the matrix
4) if rank(J) = M then the first integrals Ф1,…,Фm are functionally independent
Existence Theorem
-let |x’ = |f(|x), |xϵℝ^n and |f(|x) be a smooth vector field
THEN
in a small vicinity of any point |xo, there exist N functionally independent (time dependent) first integrals
How many time independent first integrals can be found for a particular system?
N-1
-since you can express t in terms of first integral Ф1 and x1, x2, …, xn and then sub that expression for t into the other N-1 first integrals Ф2,Ф3,…,ФN
What do we use reduction of order to do?
- having one time independent first integral, we can reduce the order of the system by one
- having k functionally independent and time independent first integrals, we can reduce the order of the system by k
How to reduce order?
1) each first integral is equal to a constant, rearrange the first integrals to write the variables in terms of these constants and one other variable e.g. y
3) substitute these expressions for the variables into the original system of differentials
4) you should now have obtained a separable first order equation in terms of y