A quantum-secured explainable artificial intelligence framework with metaheuristic optimization for scalable cloud environments

Abstract Cloud computing represents an essential platform for intelligent data processing, scalable service provision, and distributed computing structures. Unfortunately, existing cloud systems are still confronted with numerous cybersecurity issues, the absence of Explainable Artificial Intelligence (XAI), ineffective resource management, and scalability constraints, especially when it comes to new emerging quantum-based threats. Conventional cloud security instruments are based mostly on classical cryptography, meaning that they may become insecure in post-quantum conditions, while the existing approaches to optimization and explainability work separately without guaranteeing safe and scalable cloud orchestration. To overcome these problems, the research discusses the implementation of the Quantum-Secured XAI Framework with the help of Metaheuristic Optimization for scalable cloud systems. The suggested framework utilizes Quantum Key Distribution (QKD) and Post-Quantum Cryptography (PQC) for secure communication, XAI methods based on SHAP and LIME for transparent and fair AI decisions, and the Hybrid PSO-GWO algorithm for intelligent resource management and job scheduling. The experimental assessment which was performed with the aid of a number of datasets such as Google Cluster Dataset and Azure Public Datasets and simulated environment called CloudSim shows that the offered framework achieves the attack resistance amounting to 96% as well as interpretability rate equal to 94% and nearly optimal convergence results of 100% with about 17.6% superiority to GA, 6.4% to PSO, and 7.5% to GWO, besides the scalability index being equal to 98% and thus outperforming traditional methods like GA, ACO, PSO, GWO, and WOA. In addition, cross-dataset validation supports the validity of the proposed methodology in different workload conditions. The proposed approach is secure, explainable, scalable, and optimized and can be hence applied in intelligent cloud ecosystems.