Comparative Study of Radio Resource Distribution Algorithms

Abstract

The equitable distribution of radio resources among different users in wireless networks is a difficult problem and has attracted the interest of many studies. This study presents the Proportional Fair Q-Learning Algorithm (PFLA) to enable the equitable distribution of radio resources among diverse users through the integration of Q-learning and proportional fairness principles. The PFLA, Round Robin (RR), and Max Throughput (MaxTP) algorithms were compared to evaluate their effectiveness in resource allocation. Performance was measured in terms of sum-rate throughputs and fairness index. The comparison results showed an improvement in the fairness index metrics for PFLA compared to the other algorithms. PFLA showed gains of 11.62 and 43% in the fairness index compared to RR and MaxTP, respectively. These results show that PFLA is more efficient in utilizing available resources, leading to higher overall system throughput and demonstrating its ability to balance performance metrics between users, especially when the number of users increases.