زمان‌بندي بسته‌ها براي ترافيك‌ URLLC در حضور طول بلوك پويا با استفاده از الگوريتم هاي هوش مصنوعي براي شبكه هاي 6G

With the evolution of sixth-generation (6G) networks, the need for intelligent an‎d efficient resource management in the presence of heterogeneous traffic has become increasingly critical. The two primary traffic types in these networks—Ultra-Reliable Low-Latency Communications (URLLC) an‎d enhanced Mobile Broadban‎d (eMBB)—exhibit distinct an‎d sometimes conflicting requirements in terms of data rate, latency, an‎d reliability. The fundamental challenge lies in designing a packet scheduling framework capable of supporting these two services simultaneously within a dynamic an‎d complex network environment. One of the key factors in this process is the dynamic blocklength, which plays a crucial role in transmission quality, latency performance, an‎d error rate. Since classical models based on infinite blocklength assumptions fail to meet the stringent requirements of URLLC, the analysis an‎d optimization of scheduling mechanisms under the finite blocklength regime have become an essential necessity. In this research, an intelligent scheduling framework is proposed that leverages machine learning techniques—particularly deep reinforcement learning an‎d Q-learning—to make optimal resource allocation decisions under dynamic network conditions. After formulating the mathematical model of the proposed framework, the relationships among blocklength, latency, an‎d data rate are analyzed, an‎d the designed scheduling algorithm is eva‎luated in a simulation environment. The results demonstrate that integrating artificial intelligence–based algorithms with dynamic blocklength adaptation can significantly reduce URLLC latency, increase eMBB data rates, improve spectral efficiency, an‎d enhance overall system stability. Therefore, this study shows that the combination of intelligent scheduling an‎d dynamic blocklength adaptation provides a promising foundation for unified traffic management in 6G networks, offering an effective solution for achieving the desired Quality of Service (QoS) in next-generation wireless systems.

مهدي مهدوي

محمدرضا احمدزاده

حسين سعيدي , محمدرضا حيدرپور