| TEBD | |
|---|---|
| Name | Time-Evolving Block Decimation |
| Authors | Guifrém Vidal |
| Introduced | 2003 |
| Domain | Quantum physics |
| Related | Tensor network, Matrix product state, Density matrix renormalization group |
TEBD
Time-Evolving Block Decimation (TEBD) is a numerical algorithm for simulating the real- or imaginary-time evolution of one-dimensional quantum many-body systems, using a representation in terms of Matrix product states and local unitary gates. It matters in Quantum physics because it enabled efficient study of non-equilibrium dynamics, thermalization, and ground-state properties in strongly correlated systems where exact solutions are intractable. TEBD influenced later tensor network methods and practical quantum simulation tools used in both academic and industrial research.
TEBD was proposed by Guifrém Vidal in the early 2000s as a method to simulate time evolution of states represented as Matrix product states (MPS), building on ideas from the Density matrix renormalization group (DMRG) and entanglement-based truncation. The algorithm emerged amid accelerating interest in numerical methods for low-dimensional quantum systems, contemporary with improvements to DMRG by Steven R. White and conceptual advances in entanglement entropy and the role of area laws. TEBD's compact handling of local gates and singular value decompositions made it widely adopted in studies of quantum quenches, transport, and spin-chain dynamics. It has been implemented in software packages such as ITensor and contributed to collaborations at institutions like Caltech, Perimeter Institute, and Max Planck Institute for Quantum Optics.
TEBD represents a many-body wavefunction as a product of local tensors forming an MPS, exploiting low entanglement in ground states or short-time-evolved states under local Hamiltonians. Time evolution under a Hamiltonian H is approximated by a Suzuki–Trotter decomposition (a form of Trotter product formula) that factorizes the global propagator exp(-iHt) into a sequence of two-site gates. After application of each gate, a singular value decomposition (SVD) is used to restore the MPS form and truncate small Schmidt coefficients, controlling the bond dimension χ and thus the retained bipartite entanglement. The method connects to concepts from operator splitting methods and Krylov subspace approaches when higher-order decompositions or alternative integrators are used. Core theoretical considerations include the growth of entanglement entropy after quenches and the scaling behavior that determines the feasibility of long-time simulations.
A typical TEBD step alternates application of even and odd two-site gates derived from a Trotter decomposition, followed by SVD-based truncation to maintain an MPS with chosen bond dimension χ. The dominant costs are tensor contractions and SVDs, giving a per-step computational complexity scaling approximately as O(χ^3 d^3) for local Hilbert space dimension d (often d=2 for spin-1/2 systems). Memory scales as O(N χ^2 d) for a chain of length N. Practical implementations optimize gate application order, use blocking and parallel linear algebra libraries (e.g., BLAS, LAPACK), and exploit symmetries such as U(1) or SU(2) to reduce effective χ. Adaptive schemes vary χ dynamically based on discarded weight, and higher-order Trotter decompositions reduce systematic error at the cost of more gates. Variants include time-dependent variational principle (TDVP) updates that can improve stability for long times by projecting dynamics onto the MPS manifold.
TEBD has been widely applied to study dynamics and static properties in quantum spin chains, Bose–Hubbard model, and fermionic lattice models after mapping to spin variables. Key uses include simulation of quantum quenches, light-cone spreading of correlations, computation of spectral functions, and dynamics of entanglement growth. TEBD enabled detailed numerical tests of theoretical ideas such as the Eigenstate thermalization hypothesis and prethermalization phenomena. It has been used in contexts with experimental relevance, informing cold-atom experiments in optical lattices at institutions like JILA and MIT, and supporting studies of transport in low-dimensional conductors relevant to condensed matter research groups globally.
TEBD is a founding algorithm in the family of tensor network methods; it directly manipulates MPS and is philosophically interconnected with DMRG which can be viewed as a variational optimization over MPS. While DMRG excels at ground-state optimization, TEBD excels at explicit time evolution. Both methods share mathematical building blocks—SVD, Schmidt decomposition, and bond-dimension control—and have cross-fertilized improvements: time-dependent DMRG, MPS-based periodic boundary conditions treatments, and fully variational schemes like time-dependent variational principle (TDVP). TEBD and related two-site algorithms form part of the toolkit in tensor-network libraries such as TeNPy and ALPS.
TEBD's principal limitation is entanglement growth: after global quenches entanglement entropy typically grows linearly in time, forcing exponential growth of χ to retain accuracy and making long-time dynamics intractable for generic systems. Trotterization introduces systematic error controlled by timestep size and decomposition order; accumulation of truncation error also affects state fidelity. Remedies include adaptive bond-dimension schemes, higher-order Suzuki–Trotter decompositions, TDVP projection, local purification for mixed states, and employing symmetries or matrix product operator (MPO) compression. There is ongoing research integrating TEBD ideas with quantum computing primitives and with stochastic approaches to mitigate scaling challenges.
Numerical methods like TEBD are foundational in shaping which quantum problems are tractable, influencing research agendas at universities, national labs, and technology companies (e.g., IBM, Google). Equity concerns arise from concentrated access to high-performance computing and proprietary software; open-source implementations (e.g., ITensor, TeNPy) and community trainings improve accessibility for under-resourced institutions and researchers in the Global South. Ethical issues include responsible stewardship of computational resources—energy consumption of large-scale simulations—and ensuring diverse participation in methodological development so that modeling priorities reflect broader societal needs, such as materials for sustainable energy and quantum technologies that benefit public interest. Promoting open data, reproducible workflows, and inclusive collaborations aligns TEBD practice with values of scientific justice and equity.
Category:Quantum many-body physics Category:Tensor network algorithms