Metropolis Algorithms in Generalized Ensemble

TL;DR

This paper discusses generalized-ensemble algorithms, including two established methods and four new extensions, to improve simulations of complex systems like spin and protein systems by overcoming energy minima trapping.

Contribution

It introduces four novel generalized-ensemble algorithms extending multicanonical and replica-exchange methods for better sampling in complex systems.

Findings

New algorithms outperform traditional methods in complex system simulations.

Effective in overcoming energy barriers in spin and protein systems.

Validated with Potts model, Lennard-Jones fluid, and protein simulations.

Abstract

In complex systems such as spin systems and protein systems, conventional simulations in the canonical ensemble will get trapped in states of energy local minima. We employ the generalized-ensemble algorithms in order to overcome this multiple-minima problem. Two well-known generalized-ensemble algorithms, namely, multicanonical algorithm and replica-exchange method, are described. We then present four new generalized-ensemble algorithms as further extensions of the two methods. Effectiveness of the new methods are illustrated with a Potts model, Lennard-Jones fluid system, and protein system.