[Paper Review] Recent Developments in Dual Lattice Algorithms

J. Wade Cherrington|ArXiv.org|Oct 2, 2008

Complexity and Algorithms in Graphs27 references4 citations

TL;DR

This paper presents recent advances in dual lattice algorithms for non-abelian $SU(2)$ lattice gauge theories, demonstrating explicit dual amplitudes and ergodic simulation algorithms in $D=3$ and $D=4$. It shows that dynamical fermions can be naturally incorporated via a dual polymer picture, though challenges like the sign problem and critical slowing down persist in weak-coupling regimes.

ABSTRACT

We review recent progress in numerical simulations with dually transformed SU(2) LGT, starting with a discussion of explicit dual amplitudes and algorithms for SU(2) pure Yang Mills in D=3 and D=4. In the D=3 case, we discuss results that validate the dual algorithm against conventional simulations. We also review how a local, exact dynamical fermion algorithm can naturally be incorporated into the dual framework. We conclude with an outlook for this technique and a look at some of the current challenges we've encountered with this method, specifically critical slowing down and the sign problem.

Motivation & Objective

• To develop practical, ergodic simulation algorithms for non-abelian dual lattice gauge theories, particularly for $SU(2)$ in $D=3$ and $D=4$.
• To extend the dual framework to include dynamical fermions via a geometric polymer picture that couples naturally to closed branched surfaces.
• To assess the feasibility of dual simulations in weak coupling, where long auto-correlation times and the sign problem emerge as major obstacles.
• To explore the potential of dual methods for $SU(3)$, supersymmetric, and chiral fermion models in future work.

Proposed method

• The dual model is constructed by expanding the gauge action in terms of group characters and interchanging summation and integration, leading to a spin foam model with local amplitudes at plaquettes, edges, and vertices.
• Explicit dual amplitudes are derived using representation-theoretic quantities, including invariant tensors (intertwiners), enabling a local, discrete formulation of the dual theory.
• An ergodic algorithm is implemented for $SU(2)$ in $D=3$ and $D=4$, allowing numerical sampling of the dual ensemble despite non-positive definite amplitudes.
• Fermionic degrees of freedom are incorporated via a polymer expansion of the fermion determinant, which is then dualized to yield a configuration space of closed branched surfaces ending on fermion loops.
• Vertex amplitudes are modified at points where fermion charge flows through, with the modified amplitudes defined via spin network evaluations.
• The sign problem is addressed using a 'sign trick' method, though its effectiveness diminishes at weak coupling, prompting investigation into cluster algorithms and alternative sampling strategies.

Experimental results

Research questions

• RQ1 Can ergodic algorithms be constructed for non-abelian dual lattice gauge theories in $D=4$, despite complex, non-positive amplitudes?
• RQ2 How can dynamical fermions be consistently incorporated into the dual framework, and does this lead to a geometrically natural coupling to gauge degrees of freedom?
• RQ3 What is the behavior of the sign problem in $SU(2)$ dual models as the coupling is reduced toward the weak-coupling regime?
• RQ4 Can cluster algorithms or worm-type methods improve auto-correlation times in weak-coupling dual simulations?
• RQ5 To what extent can the dual formalism be generalized to $SU(3)$, supersymmetric, and chiral fermion models?

Key findings

• An ergodic algorithm was successfully implemented for $SU(2)$ lattice gauge theory in $D=3$ and $D=4$, enabling numerical validation against conventional simulations.
• Explicit dual amplitudes were derived for $SU(2)$ in both $D=3$ and $D=4$, based on representation-theoretic decomposition and invariant tensor contractions.
• The dual model naturally incorporates dynamical fermions via a polymer-like configuration space of closed branched surfaces, with modified vertex amplitudes that reflect fermion charge flow.
• The expectation value of the sign of the amplitude remained close to unity in simulations, though it decreased gradually toward weaker coupling, indicating a growing sign problem risk.
• Long auto-correlation times were observed in weak-coupling regimes, suggesting the need for advanced sampling methods such as cluster or worm algorithms.
• The dual approach remains a promising alternative for lattice gauge theory, though challenges in the sign problem and ergodicity at weak coupling remain open.