DMRG
Density Matrix Renormalization Group · Also known as: DMRG, density matrix renormalization, tensor network
Density Matrix Renormalization Group (DMRG) is a powerful computational method for solving strongly correlated quantum systems, particularly one-dimensional lattice models and quantum chemistry problems. Introduced by White in 1992, DMRG uses a variational approach and tensor-network representation to efficiently describe quantum ground states and excitations, achieving numerical accuracy competitive with exact diagonalization for systems that other methods cannot treat.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Apply DMRG to study strongly correlated quantum systems where weak-coupling perturbation theory or mean-field approaches fail: 1D quantum magnets, quantum wires, Hubbard models, and quantum chemistry in 1D or quasi-1D geometries. DMRG is optimal for systems with low entanglement. Assumes the system can be represented efficiently as an MPS.
Strengths & limitations
- Controlled approximation: bond dimension directly controls accuracy; larger bond dimensions yield more accurate results
- Scalable: treats systems 100-10,000 times larger than exact diagonalization with comparable or better accuracy
- Ground and excited states: both ground state and low-lying excited states accessible
- Real-time and imaginary-time evolution: can simulate dynamics and thermal properties
- Limited to weakly entangled systems: highly entangled 2D systems require prohibitively large bond dimensions
- Computational cost grows with bond dimension: practical systems limited to hundreds or thousands of sites
- Convergence challenges: some systems exhibit poor convergence, requiring fine-tuning of variational parameters
- Quantum chemistry: standard DMRG is 1D; extension to 2D/3D systems and periodic boundary conditions non-trivial
Frequently asked
What is the matrix product state (MPS) representation?
An MPS factorizes the many-body wavefunction as a chain of tensors: |ψ⟩ = Σ A_i₁ A_i₂ ... A_iN |i₁i₂...iN⟩. Each tensor A_iⱼ has a physical index (iⱼ) and two virtual bond indices connecting to neighbors. This representation is exact when bond dimensions are large; truncation to small bond dimension yields an efficient approximation.
How does DMRG control approximation error?
DMRG truncates the MPS bond dimension D (kept as the largest D eigenvalues of the density matrix at each step). Smaller D yields faster computation but cruder approximation; D should be chosen so that discarded eigenvalues are much smaller than kept ones. Error can be estimated from the discarded weight.
Can DMRG be applied to 2D systems?
Standard DMRG is 1D. For 2D systems, one must map the 2D lattice onto a 1D chain (e.g., snake pattern), which increases entanglement and required bond dimensions exponentially. Alternative methods (tensor network renormalization, variational projected entangled pair states) are better for 2D.
Sources
- White, S. R. (1992). Density matrix formulation for quantum renormalization groups. Physical Review Letters, 69(19), 2863-2866. DOI: 10.1103/PhysRevLett.69.2863 ↗
- Schollwöck, U. (2005). The density-matrix renormalization group in the age of matrix product states. Reviews of Modern Physics, 77(1), 259-315. DOI: 10.1103/RevModPhys.77.259 ↗
How to cite this page
ScholarGate. (2026, June 3). Density Matrix Renormalization Group. ScholarGate. https://scholargate.app/en/spectroscopy/dmrg
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
Compare side by side →