Grunnleggende konsepter
The complexity of counting subgraphs and induced subgraphs of a small pattern graph H in a large host graph G is fully classified, with dichotomies between fixed-parameter tractable and #W[1]-hard cases, when the host graph G is restricted to be somewhere dense.
Sammendrag
The paper studies the problems of counting copies (#Sub(H→G)) and induced copies (#IndSub(H→G)) of a small pattern graph H in a large host graph G. The main results present exhaustive and explicit complexity classifications for these problems when the host graph G is restricted to be a somewhere dense graph class.
Key highlights:
The problems of counting k-matchings (#Match(G)) and k-independent sets (#IndSet(G)) are identified as the minimal hard cases for #Sub(H→G) and #IndSub(H→G) respectively, when G is a monotone and somewhere dense graph class.
Dichotomy theorems are proved, showing that #Match(G) and #IndSet(G) are fixed-parameter tractable if and only if G is nowhere dense. Otherwise, they are #W[1]-hard and cannot be solved in time f(k) · |G|o(k/log k) for any function f.
These results subsume and significantly strengthen previous hardness results for counting subgraphs and induced subgraphs in restricted graph classes like bipartite graphs, F-colorable graphs, and degenerate graphs.
The proofs use a novel approach based on graph fractures and colorful tensor products, which allows for simpler and more general hardness proofs compared to previous techniques.
For homomorphism counting #Hom(H→G), a dichotomy is proved showing that the problem is FPT if G is nowhere dense or if the treewidth of H is bounded, and #W[1]-hard otherwise.