The purpose of this thesis is to study the algorithmic aspects of the small world phenomenon in large interaction networks. Experimental observations showed that large interactions networks (e.g. social or computer networks)share common global properties. One of them is the small world phenomenon which consists in the existence of very short paths between any pair of nodes, that can be discovered using only a partial knowledge of the network. We are interested in this algorithmic characteristic of the small world effect, its application to decentralized routing, and its emergence in real networks.We propose a new decentralized routing algorithm on the Kleinberg random graph model of small world, which computes paths of length O(log n.(loglo...