Starting from an existing core infrastructure capability, the network planning problem can be viewed as a resource allocation at the distribution layer. This allocation of resources is not uniform, i.e. different user profiles and use cases are considered, each associated with a quality of service. In the case of public networks, the highest priority is assigned to government institutions; guaranteeing government services is key for a city in order to improve its socioeconomic sustainability.We created an AI Tool supported by genetic algorithms together with LLM to optimize latency and bandwidth levels following the recommendations given by different standardization organizations such as ITU, IEEE and 3GPP.

G.E.N.E.O is an Application Function (AF) based on supervised genetic algorithms with language models for planning management in the insertion of Edge UPFs and their resource optimization. The planning stage optimizes the distances and resource allocation of the Edge UPFs according to the initial population distribution. In the Optimization stage, it assigns network resources according to the capacities in the Edge UPFs, considering the mobility of users for each hour. In turn, it validates the result with the 5G technical specifications given by 3GPP and generates the configuration change operation at the Edge.