{
  "id": "rayleigh-ritz-method",
  "title": "Rayleigh–Ritz Method",
  "lang": "en",
  "source_lang": "en",
  "summary": "The Rayleigh–Ritz method approximates variational problems and self-adjoint eigenvalue problems by restricting them to finite-dimensional trial spaces.",
  "category": "mathematics",
  "subcategory": "applied",
  "type": "concept",
  "aliases": [
    "Rayleigh–Ritz variational method",
    "Ritz method",
    "Ritz variational method"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/rayleigh-ritz-method/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/rayleigh-ritz-method/",
      "json": "https://aiwiki.page/zh/rayleigh-ritz-method.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "The Rayleigh–Ritz method is a technique in numerical analysis for approximating variational problems by restricting the admissible solutions to a finite-dimensional space. An unknown function or vector is expressed as a linear combination of chosen trial functions, and its coefficients are determined by minimizing an energy functional or making a Rayleigh quotient stationary. The resulting problem is a finite system of algebraic equations or a matrix eigenvalue problem. The method is central to structural mechanics, quantum mechanics, and the finite element method, and also serves as an eigenvalue extraction procedure in numerical linear algebra."
    },
    {
      "heading": "Principle and historical background",
      "text": "The method combines two closely related ideas:\n\n- Approximate a function by an element of a manageable trial space.\n- Determine the approximation through a variational condition rather than by enforcing a differential equation at every point.\n\nIts name associates the work of Lord Rayleigh on vibration frequencies and energy quotients with Walther Ritz’s systematic development of finite-dimensional variational approximations in the early twentieth century. Ritz’s work extended the approach to broad classes of boundary-value and vibration problems.\n\nTerminology varies by field. Ritz method often denotes direct minimization of an energy functional over a trial space. Rayleigh–Ritz method frequently emphasizes eigenvalue approximation through a projected operator or a stationary quotient. These are related constructions, but an energy-minimization problem need not be an eigenvalue problem.\n\nThe underlying framework belongs to the calculus of variations: a problem about functions is reduced to a problem about finitely many coefficients. Its effectiveness depends on choosing trial spaces that both satisfy the mathematical constraints and approximate the desired solution well."
    },
    {
      "heading": "Variational formulation for boundary-value problems",
      "text": "Consider a real Hilbert space \\(V\\), a continuous symmetric bilinear form \\(a:V\\times V\\to\\mathbb R\\), and a continuous linear functional \\(\\ell:V\\to\\mathbb R\\). Define the energy functional\n\n\\[J(v)=\\frac12 a(v,v)-\\ell(v).\\]\n\nSuppose \\(a\\) is coercive: there is a constant \\(\\alpha>0\\) such that\n\n\\[a(v,v)\\geq \\alpha\\|v\\|_V^2 \\qquad\\text{for every }v\\in V.\\]\n\nUnder these assumptions, \\(J\\) has a unique minimizer \\(u\\). Its first variation vanishes, giving\n\n\\[a(u,v)=\\ell(v) \\qquad\\text{for every }v\\in V.\\]\n\nThis is commonly a weak formulation of a boundary-value problem. For differential equations, it may require fewer derivatives of the solution than the corresponding pointwise equation and thereby admit a weak solution.\n\nChoose a finite-dimensional trial space\n\n\\[V_n=\\operatorname{span}\\{\\phi_1,\\ldots,\\phi_n\\}\\subset V,\\]\n\nwhere the trial functions are linearly independent. The Ritz approximation is\n\n\\[u_n=\\sum_{j=1}^{n}c_j\\phi_j,\\]\n\nwith coefficients chosen to minimize \\(J\\) over \\(V_n\\). Differentiating with respect to each coefficient gives\n\n\\[a(u_n,\\phi_i)=\\ell(\\phi_i), \\qquad i=1,\\ldots,n.\\]\n\nThus the infinite-dimensional problem becomes the linear system\n\n\\[Kc=f,\\]\n\nwhere\n\n\\[K_{ij}=a(\\phi_j,\\phi_i), \\qquad f_i=\\ell(\\phi_i).\\]\n\nThe matrix \\(K\\), often called the stiffness matrix, is symmetric positive definite under the stated assumptions. It can therefore be solved using methods such as Cholesky decomposition.\n\nBoundary conditions\n\nEssential boundary conditions, such as prescribed displacements or prescribed function values, must be built into the admissible trial functions or imposed through an appropriate constraint formulation.\n\nFor nonhomogeneous essential conditions, an admissible function \\(u_g\\) is first chosen to satisfy the prescribed boundary data, and the approximation is written as\n\n\\[u_n=u_g+\\sum_{j=1}^{n}c_j\\phi_j,\\]\n\nwhere the \\(\\phi_j\\) satisfy the corresponding homogeneous conditions. The admissible set is then affine rather than a vector space.\n\nNatural boundary conditions, such as specified fluxes or tractions, usually enter through the variational formulation after integration by parts. Trial functions need not generally satisfy those conditions pointwise."
    },
    {
      "heading": "Rayleigh–Ritz approximation of eigenvalues",
      "text": "Let \\(a\\) be a symmetric energy form and \\(b\\) a symmetric positive-definite mass or overlap form. The variational eigenproblem is\n\n\\[a(u,v)=\\lambda b(u,v) \\qquad\\text{for every }v\\in V, \\quad u\\neq 0.\\]\n\nThe associated Rayleigh quotient is\n\n\\[R(v)=\\frac{a(v,v)}{b(v,v)}, \\qquad v\\neq 0.\\]\n\nIts stationary values correspond to eigenvalues under the appropriate domain and spectral assumptions. Minimizing it gives the lowest eigenvalue when that eigenvalue exists and the variational principle applies; higher eigenvalues are characterized through additional subspace or orthogonality conditions.\n\nRestricting the problem to \\(V_n\\) produces\n\n\\[Kc=\\theta Mc,\\]\n\nwhere\n\n\\[K_{ij}=a(\\phi_j,\\phi_i), \\qquad M_{ij}=b(\\phi_j,\\phi_i).\\]\n\nThe approximate eigenvalues \\(\\theta\\) are called Ritz values, and the corresponding functions\n\n\\[u_n=\\sum_{j=1}^{n}c_j\\phi_j\\]\n\nare Ritz vectors or Ritz functions. If the basis is orthonormal with respect to \\(b\\), then \\(M=I\\); otherwise, the generalized eigenvalue problem must retain the mass matrix.\n\nMatrix formulation\n\nFor a Hermitian matrix \\(A\\), let the columns of \\(Q\\) form an orthonormal basis for a trial subspace. The reduced problem is\n\n\\[Q^*AQy=\\theta y, \\qquad x=Qy,\\]\n\nwhere \\(Q^*\\) denotes the conjugate transpose. Here \\(x\\) is a Ritz vector approximating an eigenvector of \\(A\\).\n\nIts residual\n\n\\[r=Ax-\\theta x\\]\n\nsatisfies\n\n\\[Q^*r=0.\\]\n\nThus the residual is orthogonal to the entire trial space. This makes the method a projection procedure closely related to orthogonal projection.\n\nFor a nonorthonormal basis \\(Z\\), the corresponding reduced problem is\n\n\\[Z^*AZy=\\theta Z^*Zy.\\]\n\nRayleigh–Ritz extraction can also be performed for nonsymmetric matrices, but the real-valued minimization principles and upper-bound guarantees of the self-adjoint case generally do not extend to that setting."
    },
    {
      "heading": "Variational bounds and convergence",
      "text": "Eigenvalue upper bounds\n\nFor a suitable semibounded self-adjoint operator with discrete eigenvalues, the min–max principle characterizes the ordered eigenvalues by\n\n\\[\\lambda_k= \\min_{\\substack{S\\subset V\\\\ \\dim S=k}} \\ \\max_{0\\neq v\\in S}R(v).\\]\n\nRestricting the candidate subspaces to those contained in \\(V_n\\) gives\n\n\\[\\theta_k^{(n)}= \\min_{\\substack{S\\subset V_n\\\\ \\dim S=k}} \\ \\max_{0\\neq v\\in S}R(v).\\]\n\nConsequently, a conforming Rayleigh–Ritz approximation satisfies\n\n\\[\\lambda_k\\leq\\theta_k^{(n)}, \\qquad k\\leq n.\\]\n\nIf the trial spaces are nested,\n\n\\[V_n\\subset V_{n+1},\\]\n\nthen\n\n\\[\\lambda_k \\leq\\theta_k^{(n+1)} \\leq\\theta_k^{(n)}.\\]\n\nThe approximate eigenvalues therefore decrease toward the exact ones when the spaces become sufficiently rich and the relevant spectral approximation hypotheses hold. Without nested spaces, upper bounds may still hold, but monotonicity between successive approximations is not assured.\n\nThese statements require conforming trial spaces and the correct forms. Nonconforming discretization, approximate integration, altered operators, or numerical errors can remove the upper-bound property.\n\nBest approximation in the energy norm\n\nFor the coercive source problem, subtracting the exact and discrete equations yields Galerkin orthogonality:\n\n\\[a(u-u_n,v_n)=0 \\qquad\\text{for every }v_n\\in V_n.\\]\n\nBecause \\(a\\) is symmetric and positive definite, it defines the energy norm\n\n\\[\\|v\\|_a=\\sqrt{a(v,v)}.\\]\n\nThe Ritz solution is the best approximation in this norm:\n\n\\[\\|u-u_n\\|_a = \\min_{v_n\\in V_n}\\|u-v_n\\|_a.\\]\n\nIt also satisfies the energy identity\n\n\\[J(u_n)-J(u)=\\frac12\\|u-u_n\\|_a^2.\\]\n\nIf \\(a\\) has continuity constant \\(C_a\\) and coercivity constant \\(\\alpha\\), the corresponding bound in the original space norm is\n\n\\[\\|u-u_n\\|_V \\leq \\frac{C_a}{\\alpha} \\inf_{v_n\\in V_n}\\|u-v_n\\|_V.\\]\n\nThis is a form of Céa’s lemma. Convergence follows when the trial spaces approximate the exact solution arbitrarily well in the relevant norm."
    },
    {
      "heading": "Example: the lowest eigenvalue of a fixed string",
      "text": "Consider\n\n\\[-u''(x)=\\lambda u(x), \\qquad 0<x<1,\\]\n\nwith\n\n\\[u(0)=u(1)=0.\\]\n\nThe exact eigenvalues are\n\n\\[\\lambda_k=k^2\\pi^2, \\qquad k=1,2,\\ldots.\\]\n\nThe variational setting uses the Sobolev space \\(H_0^1(0,1)\\), with\n\n\\[a(u,v)=\\int_0^1u'(x)v'(x)\\,dx, \\qquad b(u,v)=\\int_0^1u(x)v(x)\\,dx.\\]\n\nChoose the one-dimensional trial space generated by\n\n\\[\\phi(x)=x(1-x).\\]\n\nThis function satisfies both endpoint conditions. Its Rayleigh quotient is\n\n\\[R(\\phi)= \\frac{\\int_0^1(1-2x)^2\\,dx} {\\int_0^1x^2(1-x)^2\\,dx} = \\frac{1/3}{1/30} =10.\\]\n\nThe lowest Ritz value is therefore \\(10\\), an upper bound for\n\n\\[\\lambda_1=\\pi^2\\approx 9.8696.\\]\n\nThe single trial function captures the broad shape of the first vibration mode. Adding suitable independent trial functions can improve the bound. Because the quotient is unchanged by scaling \\(\\phi\\), the coefficient of a one-dimensional trial function determines normalization but not the approximate eigenvalue."
    },
    {
      "heading": "Relationship to other numerical methods",
      "text": "Galerkin methods\n\nA Galerkin method requires the residual of a weak equation to vanish against every test function in the chosen test space. For symmetric coercive problems, using the same trial and test space gives exactly the Ritz equations.\n\nThe distinction is structural: Galerkin projection applies to problems that need not arise from energy minimization, including many nonsymmetric equations. The Ritz method, in its minimization sense, relies on an underlying variational functional. A stationary functional also need not possess a minimum if the problem is indefinite.\n\nFinite elements and spectral approximations\n\nThe finite element method typically uses piecewise-polynomial trial functions with local support. For symmetric coercive problems, conforming finite element approximations are Ritz approximations. Local support leads to a sparse matrix, allowing large discretizations to be assembled and solved efficiently.\n\nGlobal polynomials, trigonometric functions, or other problem-adapted functions can instead be used as trial bases. Such choices connect Rayleigh–Ritz approximation with spectral methods and approximation theory. Smooth solutions may admit very rapid convergence, whereas singularities or limited regularity can substantially reduce the rate.\n\nIterative eigenvalue algorithms\n\nIn numerical linear algebra, Rayleigh–Ritz extraction is a component of many large-scale eigenvalue algorithms. A Krylov subspace or another adaptively constructed trial space is generated first, and a smaller projected eigenproblem is then solved within it.\n\nThe extraction procedure and the construction of the subspace are separate steps. Rayleigh–Ritz determines the approximate eigenpairs available in the current space; an iterative algorithm determines how that space is expanded, filtered, or restarted."
    },
    {
      "heading": "Applications and limitations",
      "text": "In structural mechanics, the method approximates displacement fields, equilibrium configurations, vibration modes, and, in suitable formulations, buckling loads. The matrices \\(K\\) and \\(M\\) represent stiffness and mass in vibration analysis.\n\nIn quantum mechanics, trial wavefunctions give matrix approximations to a Hamiltonian. For a self-adjoint Hamiltonian bounded below, admissible normalized trial states have energy expectations no lower than the ground-state energy. Excited-state calculations require additional variational structure, such as subspace minimization and orthogonality constraints; unrestricted minimization alone selects the lowest state available in the trial space.\n\nThe principal limitations concern approximation quality and the preservation of the variational structure:\n\n- Trial-space quality: A large space may still be ineffective if it poorly represents boundary layers, singularities, oscillations, or localization.\n- Conditioning: Nearly dependent basis functions can make the stiffness or mass matrix ill-conditioned. Basis scaling and orthogonalization improve numerical stability but do not correct an inadequate approximation space.\n- Regularity requirements: Trial functions must belong to the energy space. A formulation involving second derivatives may require more continuity than a standard second-order elliptic formulation.\n- Spectral gaps and clusters: Individual eigenvectors can be sensitive when eigenvalues are close. Approximation of the associated invariant subspace is often more informative than comparison of single vectors.\n- Spectral pollution: In some infinite-dimensional problems, especially approximation inside gaps of an essential spectrum, projected eigenvalues may converge to values outside the true spectrum. The simple low-lying discrete-eigenvalue picture does not cover these cases.\n- Computational cost: Global trial functions often yield dense matrices, while local bases generally provide more favorable sparsity.\n\nRitz values alone do not quantify every aspect of solution accuracy. Residual norms, approximation estimates, spectral separation, and checks that the discretization preserves the intended variational problem provide distinct information about the reliability of a computed result."
    }
  ],
  "links": [
    "numerical-analysis",
    "quantum-mechanics",
    "finite-element-method",
    "calculus-of-variations",
    "hilbert-space",
    "linear-functional",
    "boundary-value-problem",
    "weak-solution",
    "system-of-linear-equations",
    "cholesky-decomposition",
    "integration-by-parts",
    "conjugate-transpose",
    "orthogonal-projection",
    "self-adjoint-operator",
    "sobolev-space",
    "galerkin-method",
    "sparse-matrix",
    "numerical-linear-algebra",
    "numerical-stability"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}