Enhancing Intrusion Detection Systems with Rectified Linear Unit Function in Pigeon-Inspired Optimization Algorithm

Intrusion Detection Systems (IDS) play a pivotal role in maintaining network security by identifying and mitigating potential threats. This research introduces an innovative approach to IDS optimization by incorporating the Rectified Linear Unit (ReLU) function into the Pigeon-Inspired Optimization (PIO) algorithm. The ReLU-PIO algorithm enhances feature selection, leading to improved detection accuracy, reduced false positive rates, and faster convergence. Unlike traditional methods, the proposed algorithm leverages the dynamic adjustment capabilities of the ReLU function to refine the velocity and position updates in the optimization process, ensuring more effective exploration and exploitation of the solution space. Experimental validation using benchmark datasets, such as NSL-KDD and CICIDS2017, demonstrates the significant advantages of the ReLU-PIO approach over conventional methods, including Particle Swarm Optimization (PSO) and Genetic Algorithms (GA). The results indicate that the ReLU-PIO algorithm achieves a remarkable improvement in detection accuracy, reduces computational overhead, and lowers false alarm rates. The methodology includes a robust mathematical model, efficient implementation of ReLU-PIO, and comprehensive evaluation through various performance metrics. This work provides a transformative perspective on enhancing IDS frameworks and highlights the potential of integrating activation functions within optimization algorithms. Future research may explore the scalability of ReLU-PIO in real-time environments and its adaptability to evolving cybersecurity challenges.