Generalized cluster algorithms for frustrated spin models

Journal Article · Mon Aug 01 04:00:00 EDT 1994 · Physical Review, B: Condensed Matter; (United States)

DOI:https://doi.org/10.1103/PhysRevB.50.3058· OSTI ID:7272205

Coddington, P D ^[1]; Han, L ^[2]

---

1. Northeast Parallel Architectures Center, 111 College Place, Syracuse University, Syracuse, New York 13244 (United States)
2. Physics Department, Syracuse University, Syracuse, New York 13244 (United States)

Standard Monte Carlo cluster algorithms have proven to be very effective for many different spin models. However, they fail for frustrated spin systems. Recently, a generalized cluster algorithm was introduced that works extremely well for the fully frustrated Ising model on a square lattice by placing bonds between sites based on information from plaquettes rather than links of the lattice. Here we study some properties of this algorithm and some variants of it. We introduce a practical methodology for constructing a generalized cluster algorithm for a given spin model, and apply this method to some other frustrated Ising models. We find that such algorithms work well for simple fully frustrated Ising models in two dimensions, but appear to work poorly or not at all for more complex models such as spin glasses.

• OSTI ID: 7272205
• Journal Information: Physical Review, B: Condensed Matter; (United States), Journal Name: Physical Review, B: Condensed Matter; (United States) Vol. 50:5; ISSN PRBMDO; ISSN 0163-1829
• Country of Publication: United States
• Language: English

Related Subjects

665000* -- Physics of Condensed Matter-- (1992-)
75 CONDENSED MATTER PHYSICS
SUPERCONDUCTIVITY AND SUPERFLUIDITY
ALGORITHMS
CLUSTER EXPANSION
CRYSTAL MODELS
ISING MODEL
MATHEMATICAL LOGIC
MATHEMATICAL MODELS
SERIES EXPANSION
SPIN GLASS STATE
TEMPERATURE DEPENDENCE
TWO-DIMENSIONAL CALCULATIONS