DTMC analysis for URLLC traffic with proactive HARQ allows for accounting feedback delays and more accurately planning resources. This is critical for systems with strict SLA requirements for latency and reliability.
In URLLC systems, degradation begins not at the radio channel level, but at the planning level. Periodic streams compete for limited slots, and even a small error in estimating retransmission behavior leads to deadline violations. Proactive HARQ improves efficiency through early termination, but adds complexity: ACK arrives with a delay (HARQ RTT), and the transmitter continues to send redundant packets. Classical models ignore this effect or simplify it to a single slot, resulting in an overestimation of reliability and schedulability.
The proposed solution relies on DTMC analysis for URLLC with an expanded state space. Instead of a fixed number of retries, the model considers the probabilistic nature of successful delivery and the delay in observing ACK. This allows for the calculation of the minimum number of transmission opportunities for each packet, taking into account its deadline and required reliability. An important trade-off: the model becomes computationally more complex but more accurately reflects system behavior. A two-step approach is used to find the configuration: first, synchronous stream launches are checked, then a genetic algorithm is applied to optimize offsets.
The implementation is built around state expansion. Each packet is modeled as a set of states, including intermediate stages of waiting for ACK. HARQ RTT is explicitly encoded as a chain of sub-states through which the system deterministically passes. Transition to the next packet is only possible after reaching a state with an observable ACK. This eliminates the optimistic assumption of instantaneous feedback. The system is then analyzed slot by slot: the probability distribution of states is updated, and reliability and latency conditions are checked. If a packet reaches the target delivery probability, its state is aggregated and excluded from further analysis. On top of this model, a schedulability test and an offset optimization algorithm are built.
Results show that accounting for HARQ RTT changes the scheduling landscape. The method demonstrates higher schedulability compared to reactive HARQ and K-repetition, especially under high load. Meanwhile, the computational burden remains acceptable, although it increases in complex scenarios. Importantly, the improvement is achieved not only through the proactive HARQ mechanism but also through correct analysis of its behavior. A reduction in redundant transmissions is also observed, indicating more efficient resource utilization. Exact metrics depend on the scenario, but the initial data confirms a consistent advantage of the approach.
In an industrial context, this appears as an evolutionary improvement in scheduling models. The DTMC approach bridges the gap between theoretical guarantees and the actual behavior of 5G NR, where feedback delay and slot-level effects cannot be ignored. This is particularly important for cyber-physical systems, where planning errors can lead not only to QoS degradation but also to safety violations.
Information source
arXiv is the largest open preprint repository (since 1991, under the auspices of Cornell), where researchers quickly post working versions of papers; the materials are publicly accessible but do not undergo full peer review, so results should be considered preliminary and, where possible, checked against updated versions or peer‑reviewed journals. arxiv.org