Frequency Planning for Optimized Resource Allocation in High-Density Networks Using Machine Learning

Abstract

Frequency resource allocation in high-density networks generates significant challenges, particularly in environments with changing user demands in the midst of high congestion. Machine learning-based frequency scheduling offers dynamic resolution for optimizing network resources, such as bandwidth and frequency reuse, by using real-time data. The research work investigates the use of ML algorithms to enhance frequency reuse efficiency, focusing on maximizing network capacity while reducing interference. By integrating predictive models, adaptive frequency reuse strategies can be developed based on both historical and real-time user behavior, ensuring optimal performance during peak demand periods. This research work compares the performance of the Frequency Reuse 3 (FR3) and Frequency Reuse 7 (FR7) in high-density networks environment, emphasizing their percentage efficiency. FR3 displays a steady increase in users, peaking at 1538 users in the evening, with FR7 performing at 27.55% of FR3’s capacity during this period. In contrast, FR7 exhibits a significant evening spike to 2288 users, outperforming FR3 in handling peak traffic due to reduced interference. All through the day, FR7's shows higher percentage performance of 22.15% in the morning and a reduced rate of 10% in the afternoon, showing its relative efficiency at different times and thus better suited for peak demand, while FR3 manages steady traffic under moderate congestion.