A dichotomy for bounded degree graph homomorphisms with nonnegative weights (2002.02021v1)

Abstract: We consider the complexity of counting weighted graph homomorphisms defined by a symmetric matrix $A$. Each symmetric matrix $A$ defines a graph homomorphism function $Z_A(\cdot)$, also known as the partition function. Dyer and Greenhill [10] established a complexity dichotomy of $Z_A(\cdot)$ for symmetric ${0, 1}$-matrices $A$, and they further proved that its #P-hardness part also holds for bounded degree graphs. Bulatov and Grohe [4] extended the Dyer-Greenhill dichotomy to nonnegative symmetric matrices $A$. However, their hardness proof requires graphs of arbitrarily large degree, and whether the bounded degree part of the Dyer-Greenhill dichotomy can be extended has been an open problem for 15 years. We resolve this open problem and prove that for nonnegative symmetric $A$, either $Z_A(G)$ is in polynomial time for all graphs $G$, or it is #P-hard for bounded degree (and simple) graphs $G$. We further extend the complexity dichotomy to include nonnegative vertex weights. Additionally, we prove that the #P-hardness part of the dichotomy by Goldberg et al. [12] for $Z_A(\cdot)$ also holds for simple graphs, where $A$ is any real symmetric matrix.