AI-enabled trust-based routing protocol for social opportunistic IoT networks

The emergence of the new applications in the realm of Opportunistic IoT networks brings exceptional levels of complexity with it; hence, Artificial Intelligence-based approaches are envisioned for the design and optimization owing to the versatility and adaptability it assists in resolving complex real-time problems. In these networks, messages are transmitted through node cooperation because of the nodes' random movement and a lack of network infrastructure. As it is an open network, message forwarding is susceptible to intervention from anomalous nodes that may attribute maliciousness or selfishness, which may cause network transmission disorder and makes routing complex in Social Opportunistic IoT networks. By leveraging AI's capability, these networks can be redesigned to enable a more secure and efficient routing for the nodes. In order to suppress various harmful impacts in the network from anomalous nodes, this paper proposes an AI-Enabled Trust-Based Routing Protocol using NSGA-II for Social Opportunistic IoT Networks (TBRP), which uses the concept of multi-objective optimization. TBRP employs a new Four Tier Protocol trust scheme based on contextual parameters. This AI-enabled multi-objective optimization scheme modifies the conventional NSGA-II for amelioration through an intelligent chromosome representation, improved crossover, and mutation paradigm. TBRP is evaluated against several routing protocols and with trust-based protocols TCAFE and MT-SMRP on various performance metrics using the INFOCOMM dataset. TBRP outperforms Firefly PRoPHET, AntRouter, GAER, and PRoPHET by 1.66%, 10.66%, 25.44%, and 58.57%, respectively, in terms of the delivery success ratio while varying the time for which simulation is run.