Eigenvalue-cluster Algorithm for Matrix Monte Carlo

Abstract

Various physical models can be expressed in terms of matrices. A valuable tool for analysing matrix models is numerical simulations, often the Metropolis algorithm with various improvements. The downside of this approach is that the simulation may become stuck in a vacuous state, and probing the relevant parts of the configuration space might be difficult. Here, we propose an algorithm that moves around a cluster of eigenvalues and show that it converges to the true vacuum state.