Resource allocation and placement of virtualized functions in disaggregated radio access networks.

The optimized positioning of virtualized radio functions in the wireless access network is essential for the efficient use of resources, being a relevant research topic in 5G and Post-5G networks. This study presents three approaches to solve this problem in the network planning scenario and two approaches in the network operational scenario. A Mixed Integer Linear Programming (MILP) model is used for mathematical formulation, solved by methods based on artificial intelligence, machine learning and metaheuristics. Exact approaches provide optimal upper limits, but with low scalability, non-exact methods are more efficient and can generate high-quality solutions. The results indicate that machine learning converges quickly, while metaheuristics is more efficient in experiments. However, complexity increases in the operational landscape due to varying demand over time.