PSO-Optimized Reinforcement Learning LEACH for Wireless Sensor Networks against Black Hole and Wormhole Attacks

This research enhances the WSN method of the black hole and worm hole attackers of the LEACH using Particle Swarm Optimization and Reinforcement Learning. PDR-reported performance measures, such as end to end latency, throughput, and energy consumption, served basis for data collection in the simulations, which were carried out using the NS3 simulator. Data were divided using a time-window strategy to enhance the outcome of feature extraction and classification. The PSO was used to select the cluster head, and RL-based models were used to make the cluster adaptive and enhance the threat against various groups in a zone. The proposed hybrid PSO-RL model enhanced the network's defense against attacks by approximately 30% and 10% in terms of accuracy and F1-score in detection and minimizing false positive alarms, respectively, compared to the LEACH traditional method. The example provided enhances the design of the WSN to comply with the proposed guidelines and includes enhancements to effectively counteract adverse conditions for data transmission security.