7. Eigenvalues Flashcards
Simple Eigenvalue Problem in Matrix Form
A x = λ x
- this is satisfied for any λ when x=0
- it can only be satisfied when x≠0 for certain special values of λ, the eigenvalues λn with corresponding (non-zero) eigenvectors xn
Simplest Eigenvalue Problem for a Second-Order Differential Operator
Equation & BCs
- d²y/dx² = λy
- with y(0) = y(π) = 0
Simplest Eigenvalue Problem for a Second-Order Differential Operator
Solution
-for λ0: y = Acos(x√λ) + Bsin(x√λ) -sub in BCs, if y(0)=0 => A=0 -if y(π)=0 then for non-trivial solutions we require B≠0 so: sin(x√λ)=0 => √λ = n or λ=n² -where n is a positive integer => y = sin(nx)
Mathematical Setting
Sturm-Louisville Eigenvalue Problems
-d/dx( p(x) dy/dx) + q(x)y = λ w(x) y
-with y(a)=y(b)=0 and p(x), q(x) & w(x) given
-it is customary to consider systems with:
p(x) > 0
q(x) ≥ 0
w(x) > 0
Sturm-Louisville Eigenvalue Problems
Eigenvalues
- all eigenvalues λ are real and positive
- there are infinitely many eigenvalues, we write 0
Sturm-Louisville Eigenvalue Problems
Eigenfunctions
- eigenfunctions can be taken to be real
- eigenfunctions are orthogonal with respect to the weight function w(x)
- the eigenfunctions {yn(x)} form a complete set for functions satisfying the same boundary conditions
Sturm-Louisville Eigenvalue Problems
Orthogonality of Eigenfunctions
-eigenfunctions are orthogonal with respect to the weight function w(x):
∫ w(x) ym(x) yn(x) dx = 0, m or Im m=n
-where Im = ∫ w ym² dx
-integrals between a and b
Sturm-Louisville Eigenvalue Problems
Eigenfunctions as a Set for f(x)
f(x) = Σfj yj(x) -sum from j=1 to j=∞ => ∫ w(x) ym(x) f(x) dx = Σ ∫ w(x) ym(x) f(x) dx => fm = 1/Im ∫ w(x) ym(x) f(x) dx
Differential Operator
Simple Solution by Perturbation Analysis
-expand y = yo(x) + ε y1(x) + ε² y2(x) + …, ε«1 and sub into the equation and boundary conditions
-and specify the normalisation condition:
yo’(x) + ε y1’(x) + … = 1
-evaluate O(1) terms, a simple eigenvalue problem, there are infinitely many eigenvalue / eigenfunction solutions, take the n=1 ‘ground state’ ones
-evaluate O(ε) terms, O(ε²) terms etc.
Differential Operator
General Solution Form
-we suppose that p(x)=po(x)+εp1(x) and q(x)=qo(x)+εq1(x) with w(x) unperturbed
=>
(Lo + εL1)(yo+εy1+ε²y2+…) = w(λo+ελ1+ε²λ2+…)(yo+εy1+…)
-with:
Lo = -d/dx (po d/dx) + qo
L1 = -d/dx (p1 d/dx) + q1
-and we use y(x)=0 at x=a,b
=>
yo(x) + εy1(x) + ε²y2(x) + … = 0 at x=a,b
Differential Operator
General Solution at O(1)
Lo yo = λo w yo, yo(a)=yo(b)=0
-assume that this can be solved with eigenvalues Λj and eigenfunctions Yj(x), j=1,2,3,…
Lo Yj = Λj w Yj
-we choose yo=Ym(x) and λo=Λm
Differential Operator
General Solution at O(ε)
- take O(ε) terms, sub in yo=Ym(x) and λo=Λm from the O(1) terms
- multiply by Ym and integrate from a to b
- this allows the equation to be rewritten in terms of inner products
- rearrange for λ1
- to determine y1(x) write, y1(x)=ΣajYj(x)
- this time multiply the equation by Yj(x) and integrate from a to b to get expression for aj
Differential Operator
General Solution at O(ε²)
- take the O(ε²) terms, sub in yo=Ym(x) and λo=Λm
- multiply the equation by Ym(x) and integrate from a to b
- rearrange for λ2
Eigenvalues of Matrices
Problem
-consider:
(Ao + εA1) x = λx
-the unpertubed problem is then Ao Xj = Λj Xj for j=1,2,3,…
Eigenvalues of Matrices
Assumptions
-assume that Ao is real and symmetric so that the eigenvalues, Λj, are real
-assume that the eigenvectors are normalised and orthogonal:
Xi . Xj = δij
-assume that all Λj are distinct