Zero‐Trust‐Aware Bandwidth Allocation and Slice Isolation in 6G Networks: A Dynamic Optimization Approach

ABSTRACT 6G networks require autonomous, low‐latency, and trustworthy resource management under dynamic service demands and evolving threats. Existing frameworks treat security as a static constraint, failing to adapt to real‐time trust fluctuations and zero trust principles. We propose a zero‐trust‐aware framework for joint bandwidth allocation and slice isolation in 6G network slicing, integrating a dynamic trust evolution model, risk‐aware policy enforcement, and model‐agnostic explainability. The resource allocation problem is formulated as a constrained mixed‐integer optimization that balances service utility, security risk, and isolation costs. A computationally efficient heuristic approximation is provided for real‐time orchestration. Real‐world intrusion patterns from the CSE‐CIC‐IDS2018 dataset map to slice behaviors to ensure realism. Simulations show that the framework improves resource utilization by up to 11% over static allocation and reduces SLA violations during security incidents by 22% compared to SLA‐aware heuristics. Sensitivity analysis confirms stability under varying trust model and security weights, and the explainability module provides auditable reasoning, indicating suitability for future 6G environments.