How to Compute Spectra with Error Control.

  title={How to Compute Spectra with Error Control.},
  author={Matthew J. Colbrook and Bogdan Roman and Anders C. Hansen},
  journal={Physical review letters},
  volume={122 25},
Computing the spectra of operators is a fundamental problem in the sciences, with wide-ranging applications in condensed-matter physics, quantum mechanics and chemistry, statistical mechanics, etc. While there are algorithms that in certain cases converge to the spectrum, no general procedure is known that (a) always converges, (b) provides bounds on the errors of approximation, and (c) provides approximate eigenvectors. This may lead to incorrect simulations. It has been an open problem since… 

Figures from this paper

The computation of geometric features of spectra in the SCI hierarchy is classified, allowing us to precisely determine the boundaries of what computers can achieve and prove that the authors' algorithms are optimal.
Computing Spectral Measures and Spectral Types: New Algorithms and Classifications
It is shown that if each matrix column decays at infinity at a known asymptotic rate, then it is possible to compute spectral measures of self-adjoint and unitary linear operators on separable Hilbert spaces and is classified in the SCI hierarchy for such operators.
Pseudoergodic operators and periodic boundary conditions
A simple proof that the finite section method with periodic boundary conditions converges to the pseudospectrum of the full operator is given, and it is shown that the result carries over to pseudoergodic operators acting on lp spaces for p ∈ [1,∞].
Computing solutions of Schrödinger equations on unbounded domains- On the brink of numerical algorithms
The results provide classifications of which mathematical problems may be solved by computer assisted proofs and are a part of the Solvability Complexity Index (SCI) hierarchy and Smale's program on the foundations of computational mathematics.
Universal algorithms for computing spectra of periodic operators
It is shown that for periodic banded matrices this can be done, as well as for Schrödinger operators with periodic potentials that are sufficiently smooth, and implementable algorithms are provided, along with examples.
Computing semigroups with error control
  • M. Colbrook
  • Mathematics
    SIAM Journal on Numerical Analysis
  • 2022
It is shown that it is possible, even when only allowing pointwise evaluation of coefficients, to compute semigroups on the unbounded domain L2(Rd) that are generated by partial differential operators with polynomially bounded coefficients of locally bounded total variation.
Computing spectral properties of topological insulators without artificial truncation or supercell approximation
These tools completely avoid such artificial restrictions and allow one to probe the spectral properties of the infinitedimensional operator directly, even in the presence of material defects and disorder.
On Numerical Approximations of the Koopman Operator
The error in the Krylov subspace version of the finite section method is studied and it is indicated that Krylov sequence-based approximations can have low error without an exponential-in-dimension increase in the number of functions needed for approximation.
Spectra of Jacobi Operators via Connection Coefficient Matrices
We address the computational spectral theory of Jacobi operators that are compact perturbations of the free Jacobi operator via the asymptotic properties of a connection coefficient matrix. In
Computing eigenvalues of the Laplacian on rough domains
A general Mosco convergence theorem for bounded Euclidean domains satisfying a set of mild geometric hypotheses is proved, which implies norm-resolvent convergence for the Dirichlet Laplacian which in turn ensures spectral convergence.


Observation of parity–time symmetry in optics
One of the fundamental axioms of quantum mechanics is associated with the Hermiticity of physical observables 1 . In the case of the Hamiltonian operator, this requirement not only implies real
Undecidability of the spectral gap
This work constructs families of quantum spin systems on a two-dimensional lattice with translationally invariant, nearest-neighbour interactions, for which the spectral gap problem is undecidable, and implies that there exists no algorithm to determine whether an arbitrary model is gapped or gapless.
Making sense of non-Hermitian Hamiltonians
The Hamiltonian H specifies the energy levels and time evolution of a quantum theory. A standard axiom of quantum mechanics requires that H be Hermitian because Hermiticity guarantees that the energy
Mobility transition from ballistic to diffusive transport in non-Hermitian lattices.
It is demonstrated theoretically and experimentally that in ordered time-independent -symmetric systems, which are symmetric under space-time reflection, wave transport can undergo a sudden change from ballistic to diffusive after a specific point in time and this transition depends only on the degree of dissipation.
Lieb-Schultz-Mattis in higher dimensions
A generalization of the Lieb-Schultz-Mattis theorem to higher dimensional spin systems is shown. The physical motivation for the result is that such spin systems typically either have long-range
On the limit behaviour of second order relative spectra of self-adjoint operators
It is well known that the standard projection methods allow one to recover the whole spectrum of a bounded self-adjoint operator but they often lead to spectral pollution, i.e. to spurious
Neumann–Dirichlet maps and analysis of spectral pollution for non-self-adjoint elliptic PDEs with real essential spectrum
We prove that one of the most commonly used techniques for approximating the spectra of certain classes of non-self-adjoint elliptic partial differential equations on exterior domains does not suffer
The spectrum of a quasiperiodic Schrödinger operator
AbstractThe spectrum σ(H) of the tight binding Fibonacci Hamiltonian (Hmn=δm,n+1+δm+1,n+δm,nμv(n),v(n)= $$\chi _{[ - \omega ^3 ,\omega ^2 [} $$ ((n−1)ω), 1/ω is the golden number) is shown to
On the Solvability Complexity Index, the n-pseudospectrum and approximations of spectra of operators
It is shown that it is possible to compute spectra and pseudospectra of linear operators on separable Hilbert spaces given their matrix elements and a new classification tool for spectral problems, namely, the Solvability Complexity Index.