Velocity Trapping in the Lifted Totally Asymmetric Simple Exclusion Process and the True Self-Avoiding Random Walk

Area of Science:

• Statistical Mechanics
• Computational Physics
• Stochastic Processes

Background:

• Markov-chain Monte Carlo (MCMC) methods are crucial for simulating particle systems.
• Standard MCMC algorithms can be computationally intensive, especially for large systems.
• Achieving faster-than-physical dynamics while rigorously sampling equilibrium distributions is a key challenge.

Purpose of the Study:

• To introduce and analyze novel nonreversible MCMC algorithms for particle systems.
• To demonstrate that these algorithms can achieve sampling of the positional Boltzmann distribution.
• To investigate the dynamics and timescales of these algorithms, particularly focusing on achieving faster-than-physical rates.

Main Methods:

• Development of nonreversible Markov-chain Monte Carlo algorithms.
• Analysis of the lifted totally asymmetric simple exclusion process (lifted TASEP) as a key example.
• Investigation of velocity trapping phenomena arising from density-velocity correlations.
• Connection to many-particle realizations of true self-avoiding random walks.

Main Results:

• The proposed algorithms rigorously sample the positional Boltzmann distribution.
• These algorithms exhibit faster-than-physical dynamics due to nonthermal velocity distributions.
• Velocity trapping in lifted TASEP leads to faster out-of-equilibrium mixing and equilibrium superdiffusive timescales compared to unlifted TASEP.
• The findings are extended to models beyond one-dimensional lattices and higher dimensions.

Conclusions:

• Nonreversible MCMC algorithms offer a powerful approach for efficient simulation of particle systems.
• The lifted TASEP and similar methods provide a rigorous framework for achieving accelerated dynamics.
• The concept of velocity trapping is a key mechanism explaining the enhanced performance.
• Potential applications exist beyond traditional physics domains.