| Dirac–Frenkel variational principle | |
|---|---|
| Name | Dirac–Frenkel variational principle |
| Field | Quantum mechanics |
| Introduced | 1930s |
| Founder | P. A. M. Dirac; H. F. Frenkel |
| Notable users | John P. Boyd, Giulio Tomassetti, Hans C. Ohanian |
| Related | Time-dependent Schrödinger equation, Variational principle, Time-dependent variational principle |
Dirac–Frenkel variational principle
The Dirac–Frenkel variational principle is a time-dependent variational method used to approximate dynamics of quantum systems by projecting the full Hilbert space evolution onto a parametrized manifold of trial wavefunctions. It provides a practical route to derive equations of motion for variational parameters that respect conservation laws and symmetries of the underlying Hamiltonian, making it central in theoretical and computational Quantum mechanics for many-body and molecular dynamics.
The Dirac–Frenkel variational principle addresses the challenge of solving the time-dependent Schrödinger equation for interacting many-body systems. By restricting the state to an ansatz family, for example Gaussian wave packets, matrix product states or Hartree–Fock-type orbitals, the principle yields coupled ordinary differential equations for parameters that approximate true unitary evolution. This approach bridges foundational work of Paul Dirac on quantum dynamics with practical computational frameworks used in quantum chemistry at institutions such as the Max Planck Institute for Quantum Optics and national laboratories like Lawrence Berkeley National Laboratory.
The principle traces to ideas by Paul Dirac in the 1930s formalizing variational approaches to time-dependent problems, later articulated in tandem with work by Hermann F. Frenkel and subsequent expositors in quantum chemistry. Early adopters included researchers engaged in the development of time-dependent Hartree–Fock and Born–Oppenheimer approximation extensions. Seminal papers and textbooks by figures such as P. A. M. Dirac and later reviews in journals like Physical Review and Journal of Chemical Physics consolidated the method as a standard tool linking analytic and numerical treatments in atomic, molecular and optical physics.
Formally, given a time-dependent trial state |ψ(λ(t))⟩ parametrized by coordinates λ = {λ_i}, the Dirac–Frenkel principle requires the residual |R⟩ = iħ ∂_t|ψ⟩ − Ĥ|ψ⟩ be orthogonal to the tangent space of the variational manifold: ⟨δψ|R⟩ = 0 for all admissible variations |δψ⟩. This yields the projection condition that produces evolution equations for λ_i equivalent to stationarity of the action S = ∫ dt ⟨ψ|iħ∂_t − Ĥ|ψ⟩ under the imposed ansatz. The framework is closely related to the time-dependent variational principle (TDVP) used in derivations for multiconfigurational time-dependent Hartree (MCTDH) and Dirac notation formalism, ensuring preservation of norm and expectation values for conserved observables.
The principle underlies a broad spectrum of applications: time-dependent density functional theory (TDDFT) approximations for electronic excitations, MCTDH for molecular vibrational dynamics, and tensor-network dynamics such as time-evolving block decimation (TEBD) for lattice models. It is used in quantum control, semiclassical propagation with Gaussian wave packets, and nonadiabatic dynamics combining electronic and nuclear degrees of freedom in chemical reaction studies. Research groups in universities like University of Oxford and Massachusetts Institute of Technology employ Dirac–Frenkel-based algorithms in simulating ultrafast spectroscopy and coupled electron–phonon systems.
The Dirac–Frenkel principle is a time-dependent counterpart of stationary variational principles such as the Ritz variational method and connects to the McLachlan variational principle, which minimizes the norm of the residual. It complements the Born–Oppenheimer approximation by providing dynamics within subspaces where adiabatic separation fails. In many-body theory it dovetails with coupled cluster and configuration interaction methods when generalized to time-dependent amplitudes, and with tensor network methods like matrix product states when used as the foundation of the TDVP algorithm for one-dimensional quantum lattice systems.
Practical implementations translate projection conditions into systems of ordinary differential equations solved with standard integrators such as Runge–Kutta schemes or symplectic integrators to preserve conserved quantities. Software packages in quantum chemistry and physics—implementations in Molpro, Quantum ESPRESSO (in spirit via TDDFT), and specialized libraries for tensor network methods—embed Dirac–Frenkel-derived equations for MCTDH and TDVP workflows. Numerical stability, handling of singular metric tensors on the variational manifold, and efficient evaluation of matrix elements are central challenges addressed by algorithmic advances and high-performance computing at centers like Argonne National Laboratory.
Limitations include bias from the choice of ansatz, possible nonlinearity-induced instabilities, and difficulty capturing highly entangled dynamics outside the variational manifold. Contemporary research extends the principle via adaptive basis sets, hybrid quantum–classical schemes, stochastic variants, and error bounds informed by quantum information concepts such as entanglement entropy. Active work at institutions including California Institute of Technology and ETH Zurich focuses on combining Dirac–Frenkel approaches with machine learning ansätze and on rigorous comparisons with exact diagonalization and experimental results from platforms like quantum simulators and ultracold-atom experiments.
Category:Quantum mechanics Category:Variational methods