Social relationships and location information based routing algorithm in opportunistic networks

Abstract

In opportunistic networks, the lack of a stable route between nodes necessitates the use of relay nodes for message transmission. In social attribute-based routing, the selection of relay nodes is mainly done according to historical encounter information among nodes. This historical information is used to compute utility values between nodes, which are then used to select relay nodes. However, when there is insufficient historical encounter information between source and destination nodes, the utility value between nodes may not be accurately calculated by traditional social routing strategies. This results in difficulties in selecting appropriate relay nodes and hence a low delivery ratio. To address this shortcoming, a routing algorithm based on social relationships and location information (SRLI) is proposed. Initially, historical encounter information among nodes is used to identify social relationships among nodes. Messages are forwarded to the relay nodes possessing higher social relationships with the destination nodes. To enhance the coverage of messages, the geographic cosine similarity between nodes is calculated. Messages are preferentially transmitted to adjacent nodes with lower similarity to the current node. In addition, a dynamic buffer management strategy is utilized to facilitate the appropriate discarding of messages. Lower priority messages are prioritized for deletion when a node buffer is full. Experimental results demonstrate that the proposed routing algorithm significantly outperforms the Epidemic, Prophet, and CHOP-NET algorithms in terms of message delivery ratio, average message forwarding latency and routing overhead. Our proposed SRLI is superior to CHOP-NET, Prophet and Epidemic by at least 16.9% in terms of average message forwarding latency and 7.7% in terms of message delivery ratio, while still being able to achieve 1.5% lower routing overhead.