Research On Neighbor Discovery Algorithm In Mobile Social Networks

The explosive growth of mobile sensors and smartphones in recent years has contributed to the change of communication manner. The faster and more convenient way to communicate is advocated, which leading to the increasing attention to Mobile Social Networking. Meanwhile, the popularity of location-based services makes it possible and meaningful to associate with the Proximity-based Mobile Social Networks(PMSN). Due to existing research results are limited by mobile devices’ energy, they can not satisfy the needs of real-time in applications and the discovery will have a considerable delay.Based on the problem of neighbor discovery in mobile social networks, we conduct a deep research on bi-directional probing protocols of neighbor discovery and on-demand accelerations for indirect neighbor discovery.Firstly, we propose a neighbor discovery protocol of bi-directional probing. We apply the dynamic active time slots probing the static active time slots to accomplish the process of neighbor discovery. Then we verify the feasibility and effectiveness of the scheme in theory.According to the use of striped-probing, the worst-case discovery latency is decreased by nearly half. Meanwhile, the employment of random-probing reduces the average-case discovery latency further.Secondly, we propose a method of on-demand accelerations for indirect neighbor discovery. According to the thought of assistance by known neighbors, we can achieve the acceleration of discovering neighbor nodes. Through the calculation of temporal diversity and the spatial similarity between a known neighbor and the source node, we can obtain the slot gain at any slot between the two devices. Next, we can forecast the moving direction by the method of moving average prediction which can optimize the slot gain further. Finally, we can ensure the location of additional active time slots by the slot gain.Lastly, we use Matlab to design complete simulation experiment, which verify the performance of bi-direction probing and on-demand indirect neighbor discovery by changing the various setting of parameters.