Dynamic networks raise new knowledge discovery challenges. To handle efficiently this kind of data, an analysis method has to both decompose the network (modelled by a graph) into similar set of nodes and let the user detect structural changes in the graph. In this article we present a graph decomposition algorithm generating overlapping clusters. The complexity of this algorithmis O(|E|·deg_max^2+|V| · log(|V|))). This algorithm is particularly efficient due to its ability to detect major modifications along dynamic processes such as time related ones.