New Plain-Exponential Time Classes for Graph Homomorphism
A homomorphism from a graph G to a graph H (in this paper, both simple, undirected graphs) is a mapping f:V(G)→V(H) such that if uv∈E(G) then f(u)f(v)∈E(H). The problem Hom (G,H) of deciding whether there is a homomorphism is NP-complete, and in fact the fastest known algorithm for the general case has a running time of O *(n(H)cn(G)) (the notation O *(⋅) signifies that polynomial factors have been ignored) for a constant 0<c<1. In this paper, we consider restrictions on the graphs G and H such that the problem can be solved in plain-exponential time, i.e. in time O *(c n(G)+n(H)) for some constant c.
Previous research has identified two such restrictions. If H=K k or contains K k as a core (i.e. a homomorphically equivalent subgraph), then Hom (G,H) is the k-coloring problem, which can be solved in time O *(2n(G)) (Björklund, Husfeldt, Koivisto); and if H has treewidth at most k, then Hom (G,H) can be solved in time O *((k+3)n(G)) (Fomin, Heggernes, Kratsch). We extend these results to cases of bounded cliquewidth: if H has cliquewidth at most k, then we can count the number of homomorphisms from G to H in time O *((2k+1)max (n(G),n(H))), including the time for finding a k-expression for H. The result extends to deciding Hom (G,H) when H has a core with a k-expression, in this case with a somewhat worse running time.
If G has cliquewidth at most k, then a similar result holds, with a worse dependency on k: We are able to count Hom (G,H) in time O *((2k+1)n(G)+22kn(H)), and this also extends to when G has a core of cliquewidth at most k with a similar running time.
KeywordsGraph homomorphism Exact algorithms Cliquewidth
Unable to display preview. Download preview PDF.
- 1.Amini, O., Fomin, F.V., Saurabh, S.: Counting subgraphs via homomorphisms. In: ICALP, vol. 1, pp. 71–82 (2009) Google Scholar
- 3.Byskov, J.M.: Exact algorithms for graph colouring and exact satisfiability. PhD thesis, University of Aarhus (2005) Google Scholar
- 4.Corneil, D.G., Habib, M., Lanlignel, J.-M., Reed, B.A., Rotics, U.: Polynomial time recognition of clique-width ≤ 3 graphs (extended abstract). In: LATIN, pp. 126–134 (2000) Google Scholar
- 7.Downey, R.G., Fellows, M.: Parameterized Complexity. Monographs in Computer Science. Springer, Berlin (1999) Google Scholar
- 9.Fellows, M.R., Rosamond, F.A., Rotics, U., Szeider, S.: Clique-width minimization is NP-hard. In: STOC, pp. 354–362 (2006) Google Scholar
- 10.Flum, J., Grohe, M.: Parameterized Complexity Theory. Springer, Berlin (2006) Google Scholar
- 12.Golumbic, M.C., Rotics, U.: On the clique-width of perfect graph classes. In: WG, pp. 135–147 (1999) Google Scholar
- 13.Grohe, M.: The complexity of homomorphism and constraint satisfaction problems seen from the other side. J. ACM 54(1) (2007) Google Scholar
- 20.Jonsson, P., Nordh, G., Thapper, J.: The maximum solution problem on graphs. In: MFCS, pp. 228–239 (2007) Google Scholar
- 23.Traxler, P.: The time complexity of constraint satisfaction. In: IWPEC, pp. 190–201 (2008) Google Scholar