How Disorder Improves Navigation in Dynamic Networks

Dynamic networks like Online Social Network or Disruption Tolerant Network exhibit some degree of regularity on their temporal contact patterns. The impact of this regularity on network performances has not been well studied and analyzed. In this paper, we study this temporal dimension of dynamic networks and its impacts on routing performances. We propose a simple parametric network model which covers the full spectrum of contact patterns from strictly periodic to fully random. Based on this model, we study the impact of contact patterns regularity on routing performance and we show how to exploit the temporal structure to navigate with a good resource/performance tradeoff in a dynamic network. Simulation and analytical analysis show that efficient navigation with respect to their degree of regularity emerge within a subset of the dynamic networks. Moreover, we show there is a degree of regularity where navigation performance achieves its optimum.