Research on Optimization of Network Mass Event Preplanning Based on Machine Learning

Effective management of 5G resources requires careful arranging of working numerology lists and range conveyance as per the necessities of the ongoing traffic. To lessen signal above, any reconfigurations made to oblige the ongoing traffic design should be kept to a base. This paper proposes the pre-arranging of numerology profiles as a solution for this issue and fosters a numerical improvement model for their pre-arranging. By looking at the prerequisites and limitations that are much of the time present in a specific remote correspondence climate, it is expected to make numerology profiles. One of the profiles will then, at that point, be utilized in accordance with the ongoing clients and traffic design. The model empowers the enhancement of blended numerologies in future 5G organizations under any remote correspondence situation, with explicit assistance necessities and weaknesses, and under any traffic situation. By involving attainable rate execution as one of the key qualities, we look at the blended numerology and small scale space methods in different settings. We employ Cyclic Prefix-Orthogonal Frequency Division Multiplexing (CP-OFDM) as a suitable benchmark and show that Universal Filtered Multicarrier (UFMC) outperforms CP-OFDM. When compared to a non-optimized technique, the proposed optimisation model can achieve satisfaction levels of 60–100% while reducing the overall number of numerology indexes in use.