Math 2.0

Calculus of variations and optimal control; optimization Definition

Rayleigh–Ritz method

Builds on Hilbert spaceSelf-adjoint operatorRayleigh quotientEigenvalueMin-max theorem

Definition

Let HH be a real or complex Hilbert space with inner product ⟨⋅,⋅⟩\ip{\cdot}{\cdot}. Let AA be a self-adjoint operator in HH that is bounded below. Let QQ be its closed quadratic form with form domain D(Q)D(Q), so that Q(u,v)=⟨Au,v⟩Q(u, v) = \ip{Au}{v} for uu in the domain of AA. Let m≥1m \ge 1 and let φ1,…,φm∈D(Q)\varphi_1, \dots, \varphi_m \in D(Q) be linearly independent trial vectors. Let KK and MM be the m×mm \times m matrices with entries Kij=Q(φj,φi)K_{ij} = Q(\varphi_j, \varphi_i) and Mij=⟨φj,φi⟩M_{ij} = \ip{\varphi_j}{\varphi_i}. The Rayleigh–Ritz method computes the solutions of the generalised eigenvalue problem Kc=μMc,c∈ℂm∖{0}.K c = \mu M c, \qquad c \in \C^m \setminus \set{0}. Its eigenvalues μ1≤⋯≤μm\mu_1 \le \cdots \le \mu_m are the Ritz values, and the vectors ∑jcjφj\sum_j c_j \varphi_j for the eigenvectors cc are the Ritz vectors.

Remarks

The Ritz values are the stationary values of the Rayleigh quotient Q(u,u)/⟨u,u⟩Q(u, u) / \ip{u}{u} on the span VV of the trial vectors. If the spectrum of AA below the essential spectrum consists of eigenvalues λ1≤λ2≤⋯\lambda_1 \le \lambda_2 \le \cdots, the min-max theorem gives λk≤μk\lambda_k \le \mu_k for k=1,…,mk = 1, \dots, m. For a Hermitian n×nn \times n matrix AA and a matrix WW with mm orthonormal columns, the Ritz values are the eigenvalues of W∗AWW^* A W.

Example: H=L2(0,1)H = L^2(0, 1) and A=−d2dx2A = -\frac{d^2}{dx^2} with Dirichlet conditions, so D(Q)=H01(0,1)D(Q) = H^1_0(0, 1), Q(u,u)=∫01∣u′∣2 dxQ(u, u) = \int_0^1 \abs{u'}^2 \, dx and λ1=π2\lambda_1 = \pi^2. With m=1m = 1 and φ1(x)=x(1−x)\varphi_1(x) = x (1 - x), which vanishes at 00 and 11, one gets K=∫01(1−2x)2 dx=13K = \int_0^1 (1 - 2x)^2 \, dx = \frac{1}{3} and M=∫01x2(1−x)2 dx=130M = \int_0^1 x^2 (1 - x)^2 \, dx = \frac{1}{30}. So μ1=10\mu_1 = 10, an upper bound for π2≈9.8696\pi^2 \approx 9.8696.

Non-example: the trial function φ1=1\varphi_1 = 1 is not in H01(0,1)H^1_0(0, 1), because it does not vanish at the ends. Its Rayleigh quotient is 0<π20 < \pi^2, so the upper bound fails when the trial vectors leave the form domain.

Reading path

To understand this page from the ground up, read these 8 pages in order. The foundations come first.

Show the path as a graph
Each page uses the pages above it. The colour of the bar on a box shows how strong the evidence is, and the box also gives it as text. The list has the same pages.
  • no evidence level
  1. Dirichlet eigenvalue concept
  2. Eigenvalue concept
  3. Hilbert space concept
  4. Linear subspace concept
  5. Rayleigh quotient concept
  6. Self-adjoint operator concept
  7. Min-max theorem statement
  8. Rayleigh–Ritz method definition

Further reading

  1. Web https://en.wikipedia.org/wiki/Rayleigh%E2%80%93Ritz_method
About this page
TagBAGCJR5EQF
URIm2://BAGCJR5EQF
Kinddefinition
WritingW0 stub
MSC49R05, 65N25, 65F15
DefinesRayleigh–Ritz method
UsesHilbert space, Self-adjoint operator, Rayleigh quotient, Eigenvalue, Min-max theorem

Markdown · JSON · Edit on GitHub

What this page needs for W1

  • an inventory row: the row of the inventory that the node accounts for (0 of 1)
  • a source with a locator (0 of 1)

The writing levels say how complete the text of a page is. They do not measure its truth.

What links here

No page links to this page yet.