Throughput maximization and overhead-aware quality link scheduling for heterogeneous wireless networks

Abstract

Next-generation heterogeneous wireless networks (HWNs) require advanced scheduling solutions to guarantee high quality of service (QoS) while sustaining throughput, minimizing latency, and efficiently utilizing spectrum resources. Traditional link scheduling and admission control techniques are constrained by interference complexities and dynamic mobility patterns, leading to suboptimal connectivity and reduced network performance. To address these limitations, this paper introduces a throughput maximization and link overhead minimization-aware scheduling (TMLOMS) strategy. The proposed mechanism embeds adaptive backoff-time optimization into resource allocation and incorporates a priority-driven call admission control process powered by an optimized deep reinforcement learning (ODRL) framework. The TMLOMS-ODRL is mathematically modeled to balance trade-offs between spectral efficiency, throughput maximization, and latency minimization, thereby enhancing QoS across diverse user requirements. Performance is validated using the Stanford University Interim (SUI) channel model under both urban and highway propagation environments, capturing realistic interference and mobility dynamics. Comparative evaluation against the hybrid beamforming (HBF) optimized through DRL method demonstrates that TMLOMS-ODRL significantly reduces resource access failures, lowers delay, and improves delivery ratio and throughput. The proposed strategy establishes a scalable and intelligent link scheduling mechanism for HWNs, ensuring resilient QoS-driven performance under complex real-world conditions.