Counting independent sets via Divide Measure and Conquer method

In this paper we give an algorithm for counting the number of all independent sets in a given graph which works in time $O^*(1.1394^n)$ for subcubic graphs and in time $O^*(1.2369^n)$ for general graphs, where $n$ is the number of vertices in the instance graph, and polynomial space. The result comes from combining two well known methods "Divide and Conquer" and "Measure and Conquer". We introduce this new concept of Divide, Measure and Conquer method and expect it will find applications in other problems. The algorithm of Bj\"orklund, Husfeldt and Koivisto for graph colouring with our algorithm used as a subroutine has complexity $O^*(2.2369^n)$ and is currently the fastest graph colouring algorithm in polynomial space.