Inter-node relationships in short-range mobile social networks

摘要

Empirical studies have provided evidence suggesting that human mobility and the resulting contact opportunities of short-range mobile social networks are not entirely random. Most of these studies claim that power-law characterises diverse aspects of human mobility patterns. However, is the power-law distribution assumption really reliable? Furthermore, these studies typically focus on aggregated statistics or on models for individual users, and do not capture the inter-node relationships. We empirically studied some statistical properties of internode relationships, and proposed a universal scheme, named A-Scheme, to improve existing probabilistic routing approaches. In most of those approaches, decision is made using a static method. However, we claim that judging threshold varies according to activity is more suitable. Simulation results show that A-Scheme improves the performance of existing probabilistic routing approaches in terms of cost, success rate and delay.