(IoT) Network intrusion detection system using optimization algorithms.

Area of Science:

• Cybersecurity
• Artificial Intelligence
• Swarm Intelligence

Background:

• Internet of Things (IoT) environments face complex network intrusion detection challenges.
• Existing methods often struggle with large feature spaces and slow convergence.
• Single optimization algorithms are susceptible to local optima.

Purpose of the Study:

• To propose a hybrid intelligent framework (WOA-GWO) for enhanced IoT network intrusion detection.
• To balance global exploration and local exploitation for improved optimization.
• To overcome limitations of single-algorithm approaches in detecting diverse cyber threats.

Main Methods:

• Integration of Whale Optimization Algorithm (WOA) and Grey Wolf Optimization (GWO) into a hybrid WOA-GWO framework.
• Leveraging WOA's spiral bubble-net strategy for global optimization and GWO's social hierarchy for local exploitation.
• Implementing dynamic feature selection and parameter optimization for improved detection sensitivity and generalization.

Main Results:

• The WOA-GWO framework demonstrated superior sensitivity and generalization compared to LSTM-RNN and SVM.
• Achieved superior real-time responsiveness in binary classification tasks due to a lightweight design.
• Effectively mitigated feature confusion in multi-class attack identification through adaptive feature weight allocation.

Conclusions:

• The hybrid WOA-GWO algorithm offers a novel methodological foundation for efficient threat detection in IoT security.
• Swarm intelligence algorithms show significant potential for securing resource-constrained IoT environments.
• The proposed framework provides a robust solution for dynamic feature selection and parameter optimization in intrusion detection.