Development and Simulation of an Algorithm for Intelligent Radio Resource Allocation Based on the Open RAN Architecture xApp and rApp Components Interaction

Abstract

This article examines the problem of intelligent radio resource allocation in radio access networks built on the principles of the O-RAN open architecture. It is shown that the scheduling algorithms used in existing systems, which focus on one or two target metrics, are unable to fully account for the simultaneous changes in multiple network parameters, including radio channel quality, traffic intensity, and user latency and packet loss requirements. As a solution, a model for interaction between xApp and rApp components is proposed. It implements a two-level control scheme: a long-term network state forecast generated in the Non-RT RIC loop and operational adjustments to the distribution of physical resource blocks performed in the near-RT RIC loop. The mathematical apparatus of the model was developed, including a formalization of the resource allocation problem taking into account the limitations of OFDMA technology, a proprietary integrated user prioritization criterion combining channel quality, service priority, load forecast, and service history, and a stability analysis of the proposed control algorithm based on the Lyapunov function method. The model's performance was verified through simulation in the MATLAB Simulink environment, reproducing the full system operation cycle - from user traffic generation to the generation of control actions on the base station. The results demonstrate reduced data transmission latency, a reduced probability of packet loss, increased spectral efficiency, and increased fairness in resource distribution among users compared to traditional scheduling algorithms.