Abstract:We discuss the development of cluster algorithms from the viewpoint of probability theory and not from the usual viewpoint of a particular model. By using the perspective of probability theory, we detail the nature of a cluster algorithm, make explicit the assumptions embodied in all clusters of which we are aware, and define the construction of free cluster algorithms. We also illustrate these procedures by rederiving the Swendsen-Wang algorithm, presenting the details of the loop algorithm for a worldline simulation of a quantum $S=$ 1/2 model, and proposing a free cluster version of the Swendsen-Wang replica method for the random Ising model. How the principle of maximum entropy might be used to aid the construction of cluster algorithms is also discussed.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to develop and understand cluster algorithms for physical system simulations, especially from the perspective of probability theory rather than from the perspective of specific models. Specifically, the goals of the paper include: 1. **Describe in detail the essence of cluster algorithms**: Explain how cluster algorithms work from the perspective of probability theory and clarify the implicit assumptions in all existing cluster algorithms. 2. **Define free cluster algorithms**: Propose a new type of cluster algorithm - free cluster algorithms, in which the interactions between clusters disappear, thus simplifying the design and implementation of the algorithm. 3. **Re - derive and improve existing algorithms**: By re - deriving the Swendsen - Wang algorithm, the loop algorithm of the quantum S = 1/2 model in world - line simulations, and the Swendsen - Wang replica method of the random Ising model, show how to use the new framework to improve existing algorithms. 4. **Explore the application of the maximum entropy principle**: Discuss how to use the maximum entropy principle to assist in the construction of cluster algorithms to improve the efficiency and applicability of the algorithms. 5. **Provide a more general cluster algorithm framework**: Establish a cluster algorithm framework that can be applied to a wider range of systems, especially when dealing with quantum systems, to overcome the unnaturalness of the concepts of effective Hamiltonian and local interaction energy in classical systems in quantum systems. ### Summary of main problems The core problems of this paper are: - How to understand and develop cluster algorithms from the perspective of probability theory. - How to design free cluster algorithms so that the interactions between clusters disappear, thereby simplifying calculations and improving efficiency. - How to use the maximum entropy principle to optimize the selection and construction of cluster algorithms. By solving these problems, the paper aims to provide more efficient and more general algorithmic tools for Monte Carlo simulations of physical systems.

Dual Monte Carlo and Cluster Algorithms

Applications of Cluster Perturbation Theory Using Quantum Monte Carlo Data

Cluster-Algorithm-Amenable Models of Gauge Fields and Matter

Cluster Monte-Carlo Algorithm For The Study Of Quantum X-Y Model

Cluster Algorithm for hard spheres and related systems

Single-cluster dynamics for the random-cluster model

Cluster variation method in statistical physics and probabilistic graphical models

Cluster Algebras: Network Science and Machine Learning

Critical dynamics of cluster algorithms in the random-bond Ising model

Large scale parallelization in stochastic coupled cluster

Geometric allocation approaches in Markov chain Monte Carlo

Sweeping cluster algorithm for quantum spin systems with strong geometric restrictions

An efficient cluster algorithm and the study for the two-dimensional quantum heisenberg antiferromagnet

Optimal Clustering under Uncertainty

A parallelizable model-based approach for marginal and multivariate clustering

Multigrid Monte Carlo Revisited: Theory and Bayesian Inference

Meron-Cluster Algorithms for Quantum Link Models

Exact Calculation of the Probabilities of Rare Events in Cluster-Cluster Aggregation

Cluster variation method and disorder varieties of two-dimensional Ising-like models

Parallelized event chain algorithm for dense hard sphere and polymer systems

Cluster dynamics modelling of materials: A new hybrid deterministic/stochastic coupling approach