Proposed by Ciaran McCreesh
Given a simple undirected graph $G = (V,E)$, where $V$ is the set of vertices and $E$ the set of undirected edges, a clique is a subset of $V$ such that each distinct pair of vertices in this subset are adjacent. The maximum clique problem is to find a clique of largest cardinality within a given graph. (The related clique enumeration problem is to enumerate all maximal cliques–that is, cliques which cannot be extended by adding an additional vertex.)
The second DIMACS implementation challenge studied this problem, and provided a standard set of benchmark instances in a simple file format. These instances are of varying size and difficulty: some should be trivial, but a few are still open. A simple example of a file is:
c Lines that start with a c are comments. There is one line at the start c which starts with either "p edge" or "p col", followed by the number of c vertices and the (possibly incorrect) number of edges. The e lines each c describe an edge. Some files have blank lines. p edge 5 6 e 1 2 e 2 3 e 3 4 e 4 1 e 3 5 e 4 5
This describes a graph with 5 vertices (which are numbered 1 to 5) and 6 edges. The edges number is not reliable and should be ignored. Some instances include (some) edges in both directions, whilst others do not. Some files contain loops (vertices adjacent to themselves), which should be ignored for the clique problem.
The maximum clique in this file has size 3, with vertices 3, 4 and 5.
Other datasets are available, also using this format.
The maximum clique problem is equivalent to the maximum independent set problem and the vertex cover problem. It is also used as an intermediate step in solving the maximum common subgraph problem.