Theoretical and Empirical Investigation of Artificial Intelligence-Driven Network Slicing for Healthcare Telecommunications Services in Nigeria

Network slicing in 5G systems can create dedicated virtual network instances with guaranteed resource isolation and quality of service customization for healthcare telecommunications services including telemedicine, patient monitoring, and emergency medical communication, addressing the reliability and latency requirements that shared public network slices cannot provide. The theoretical foundations and empirical validation of AI-driven dynamic network slicing optimized for healthcare telecommunications in the specific service and population context of Nigeria represent a doctoral research contribution of direct national health system relevance. This doctoral study develops an original theoretical framework and empirical validation for AI-driven network slicing for healthcare telecommunications services across Nigerian geopolitical zones. The theoretical framework formalizes healthcare slice resource allocation as a constrained optimization problem incorporating stochastic service demand models derived from empirical telemedicine traffic characterization studies conducted in partnership with three Nigerian tertiary hospitals. An original graph neural network architecture for predictive slice resource allocation is developed, learning the spatiotemporal demand patterns of healthcare telecommunications traffic to enable proactive resource provisioning before demand peaks. The GNN architecture's sample complexity and generalization bounds under healthcare traffic distributional assumptions are analytically characterized, constituting an original contribution to learning theory for network management. Experimental validation using an Open RAN testbed with slice emulation demonstrates that the AI-driven framework maintains latency below the 100 ms telemedicine requirement at 99.4% of time steps, compared to 92.7% for static slice allocation under measured demand variability. Keywords: network slicing, AI-driven network management, healthcare telecommunications, 5G, Nigeria.