Human mobility models and routing protocols for mobile social networks

In mobile social networks (MSNs), there are a number of challenges to face, which include the design of well-performed routing protocols. Since performance evaluation of a routing protocol in real world is costly and time consuming, it is usually more practical to evaluate the performance relying on the synthetic data generated by simulations. Accordingly, a number of mobility models have been proposed to provide the real-trace-like scenarios that can be used in development and performance evaluation of routing protocols. In this thesis, we concern the problem of dynamic routing in MSNs. First, two mobility models are proposed, which are the Preferred-Community-Aware Mobility (PCAM) and Role Playing Mobility (RPM) models. While designed based on thesimplicity and randomness of the Random WayPoint (RWP) mobility model, the PCAM model enhances it by exploiting the fact that people often have their favorite places to visit, resulting in the so-called human social behavior of mutual preferred communities. On the other hand, the RPM model further improves the PCAM model by jointly considering people’s behaviour of mutual-preferred-communities and their daily schedules. Then, based on the mobility models proposed, we design some routing protocols and investigate the effect of human social behavior on the routing performance in MSNs. Firstly, a Social Contact Probability assisted Routing (SCPR) protocol is proposed, which is capable of exploiting the properties of encounters between mobile nodes (MNs) and the relationship strength between MNs and communities. Secondly, considering that the energy-efficiency issues have not been addressed in the existing routing protocols for MSNs, we propose two types of energy-concerned routing protocols, which are the Energy-Concerned Routing (EnCoR) and Energy-efficient Social Distance Routing (ESDR) protocols. It can be shown that these two types of routing protocols are capable of achieving the trade-off among Energy Consumption (EC), Delivery Ratio (DR), and Delay (D), shortened as the EC/DR/D trade-off. Specifically, the EnCoR protocol controls the EC/DR/D trade-off by a threshold introduced to the route selection process, while the ESDR protocol achieves the EC/DR/D trade-off by taking into account of the length of messages and the estimation of the number of hops. Furthermore, we show that, by invoking our EnCoR or ESDR scheme, most existing routing protocols can be readily extended to their corresponding versions that are flexible to achieve the EC/DR/D trade-off. In this thesis, the performance of the proposed routing protocols are investigated with the aid of simulations, which are compared with a range of existing routing protocols for MSNs. Our studies and performance results show that the proposed protocols are capable of efficiently integrating the merits of high delivery ratio, low delivery latency and low resource consumption of the existing protocols, while simultaneously circumventing their respective shortcomings. Our proposed protocols are capable of attaining a good trade-off among the delivery ratio, average delivery delay, and the cost of resources (including energy) for operation of the protocols