Mobile opportunistic networks (MONs) are deliberated as important aspect to proliferate the wireless communications. These networks pose several challenges such as network lifetime, storage capacity, and forwarding capacity. End-to-End routing schemes are considered as promising technique solution to deal with these issues. In this domain, opportunistic routing has gained huge attention because it follows the broadcasting nature of wireless communication and focus on selection of relay node for packet transmission to ensure the better quality of service (QoS) and energy efficiency. This work focuses on optimizing the next hop selection process and introduced reinforcement learning approach which considers distance, energy and link connectivity to assign the reward for different actions to identify the suitable relay node. Moreover, geo-context and social behaviour based opportunistic routing models are used to increase the reliability of next hop selection. Similarly, social model considers social profiling, social connectivity, and social interaction model to identify the relay node. The outcome of proposed approach is compared with several existing approaches such as prophet, spray and wait, and epidemic routing in terms of packet delivery, and network overhead. The relative study shows that the proposed approach achieves the average packet delivery as 47.22% and minimizes the network overhead.
Read full abstract