OPTIMIZATION-DRIVEN INTRUSION DETECTION SYSTEMS: A SYSTEMATIC REVIEW OF ALGORITHMS, FRAMEWORKS, AND ADVANCEMENTS

Abstract

The rapid growth of IoT and Vehicle-to-Grid (V2G) systems has exacerbated cybersecurity vulnerabilities and therefore Intrusion Detection Systems (IDS) need to be able to function in highly dynamic, resource constrained, and het-erogeneous environments. Recent studies focus more on the integration of optimization algorithms in order to improve the accuracy of the IDS, as well as to minimise false alarms and enhance adaptability. However, the current research is fragmented, and there is currently no consolidated understanding of the role of optimization driven techniques in robust threat detection in these interconnected areas. This systematic review summarizes the progress in recent years, including swarm intelligence, evolutionary algorithms, multi-objective optimizers, and hybrid learning-optimization frameworks to be de-ployed in IoT and V2G environments. Comparative analysis shows that op-timization has a dramatic improvement on the performance of IDS - espe-cially in feature selection, threshold optimization, classifier optimization - but there are serious issues in scalability, real-time response and cross-domain transferability. Through the mapping of methodological trends, the identification of limitations and a discussion of the future, this review can offer a comprehensive basis for designing next-generation IDS architectures. The contribution of the study is to provide a common point of view on opti-mization-driven IDS research, which will serve to develop more resilient, adaptive, and deployable solutions for emerging cyber-physical ecosystems.