[Paper Review] Real Space Renormalization Group Techniques and Applications

Javier Rodríguez-Laguna|arXiv (Cornell University)|Jul 13, 2002

Random Matrices and Applications3 references3 citations

TL;DR

This PhD thesis introduces novel real-space renormalization group (RSRG) techniques—specifically the Correlated Blocks Renormalization Group (CBRG) and Puncture Renormalization Group (PRG)—to study quantum many-body systems in one and higher dimensions. It combines variational methods, density matrix renormalization group (DMRG) algorithms, and discrete field evolution to achieve accurate ground state and dynamics calculations, with key results showing convergence to exact solutions and successful application to disordered excitonic systems and dendrimers.

ABSTRACT

Real Space Renormalization Group (RSRG) techniques and their applications, mainly to quantum mechanics and to partial differential equations, are discussed. Special emphasis is given to the theoretical insight and the reasons for the success of some techniques, specially DMRG. Applications to the spectrum of dendrimers and excitons on random media are considered. This work is the Ph.D. of the author. Almost all the results have already been published. The main interest of this text is paedagogical and as a review.

Motivation & Objective

• To develop real-space renormalization group techniques for accurate simulation of quantum many-body systems.
• To extend DMRG algorithms to long-range and multidimensional systems using novel formulations.
• To apply RSRG to field evolution equations with physical truncation criteria.
• To enable efficient, exact diagonalization and wavefunction reconstruction in complex quantum systems.
• To provide open-source computational tools for scientific reproducibility and pedagogical use.

Proposed method

• Proposes the Correlated Blocks Renormalization Group (CBRG) for 1D and 2D systems using block-based coarse-graining with correlated truncation.
• Develops the Puncture Renormalization Group (PRG) for long-range and multidimensional lattices via punctured lattice renormalization and blocks algebra.
• Adapts the DMRG algorithm to 1D potentials, trees, and dendrimers, including exciton dynamics with disorder.
• Employs exact diagonalization via Householder and QL algorithms for accurate low-energy spectrum computation.
• Uses discrete heat equation asymptotics to validate Gaussian spreading of initial delta functions under diffusion.
• Implements custom C++ libraries for matrices, graphs, operators, and X11/PostScript graphics for full computational reproducibility.

Experimental results

Research questions

• RQ1 How can real-space renormalization group techniques be systematically extended to correlated and long-range interacting systems?
• RQ2 What is the role of block correlation and self-replicability in improving convergence of variational RG methods?
• RQ3 Can DMRG be generalized to non-1D topologies such as trees and dendrimers with physical accuracy?
• RQ4 How do discrete field evolution equations reproduce continuum diffusion behavior under finite-difference schemes?
• RQ5 What physical criteria can guide optimal truncation in real-space RG without relying on ad hoc choices?

Key findings

• The CBRG method achieves convergence to exact ground states in 1D quantum models, validated through wavefunction reconstruction and energy minimization.
• The discrete heat equation with finite differences reproduces the continuum Gaussian spreading behavior, with second moment growing linearly as ⟨x²⟩(t) = ⟨x²⟩(0) + 2κmt.
• The Puncture Renormalization Group (PRG) successfully generalizes DMRG to 2D and 3D lattices using punctured lattice renormalization and blocks algebra.
• The DMRG algorithm applied to dendrimers accurately models exciton dynamics, including effects of disorder in energy transfer.
• The method achieves exact diagonalization of Hamiltonians via Householder tridiagonalization and QL algorithm with implicit shifts, ensuring high numerical precision.
• All computational tools, including libraries for graphs, matrices, and graphics, are released under the GPL, enabling full reproducibility and reuse.