AI-Driven Swarm-Optimized Adaptive Routing Using Quantum-Inspired Neural Scheduling with Homomorphic Encryption

Abstract

The fast increase in network traffic and the moving nature of nodes in modern communication systems have led to the need to have intelligent ways of analysing traffic, managing clusters, efficiently routing, and ensuring safety in the transmission of data. Conventional approaches are not always able to deal with the complexity and size of multi-criteria network settings. This paper introduces a new multi-phase intelligent network management framework which combines deep learning, evolutionary optimization, and a quantum-inspired algorithm to improve performance, reliability, and security. The initial step uses the Hybrid Autoencoder-GAN Behaviour Synthesizer (HAE-GANBS) to examine traffic data of the Multi-Criteria Network Routing Dataset, recreate typical traffic, and create synthetic flows, which augment the feature description. The enhanced dataset is used as input into the Hybrid Spiking Neural-Evolutionary Cluster Leader Selector (HSN-ECLS) which determines the best cluster leaders through temporal spike-train modelling and multi-criteria fitness assessment. Predictive Evolutionary Trust-Aware Scheduler and Router (PETASR) is a predictive scheduling based on evolutionary operations to schedule routing paths based on future traffic, node availability, and trustworthiness. Lastly, the Quantum-Inspired Neural Scheduler-Router with Homomorphic Encryption (QINSR-HE) ensures the safety of information transfer, providing the opportunity to use encrypted, adaptive, and trust-conscious routing. The evaluation of performance indicates that the framework is more efficient in large, volatile network systems due to enhanced traffic predictability, cluster stability, routing competence, and secure information transfer over the network.