Abstract
Ultra-reliability and low latency communication has long been an important but challenging task in the fifth and sixth generation wireless communication systems. Scheduling as many users as possible to serve on the limited time-frequency resource is one of a crucial topic, subjecting to the maximum allowable transmission power and the minimum rate requirement of each user. We address it by proposing a mixed integer programming model, with the goal of maximizing the set cardinality of users instead of maximizing the system sum rate or energy efficiency. Mathematical transformations and successive convex approximation are combined to solve the complex optimization problem. Numerical results show that the proposed method achieves a considerable performance compared with exhaustive search method, but with lower computational complexity.
1. Introduction
In the evolution toward 6G networks, Ultra-Reliable Low-Latency Communication (uRLLC) has emerged as a critical pillar for applications such as industrial automation and autonomous driving. Unlike traditional mobile broadband which prioritizes aggregate throughput (sum rate), uRLLC environments are characterized by strict reliability (e.g., $10^{-5}$ to $10^{-9}$ error rates) and sub-1ms latency constraints. In ultra-dense scenarios, the primary challenge shifts from maximizing individual data rates to maximizing the number of users that can be simultaneously served within limited spectral and power resources. This "set cardinality" problem is inherently complex due to its mixed-integer programming nature, which is typically NP-hard.
2. Methodology
The proposed approach addresses the joint optimization of User Scheduling (US) and Beamforming (BF) under Finite Blocklength (FBL) constraints. Standard Shannon capacity models are often too optimistic for short-packet uRLLC transmissions; thus, FBL theory is employed to provide a realistic performance bound.
To solve the non-convex cardinality maximization problem, the authors introduce binary indicator variables $\kappa_k$ for user scheduling. These discrete variables are relaxed into a continuous range $[0, 1]$ with a penalty function to enforce sparsity and push values toward binary outcomes. The resulting non-convex problem is then solved using Successive Convex Approximation (SCA), which iteratively linearizes the objective and constraints into a series of solvable convex sub-problems. A "tuning" post-processing step is further introduced to re-allocate residual power, potentially increasing the number of scheduled users.
The simulation environment considers a single base station with 4 antennas serving up to 18 users within a 300m radius, modeled with Rayleigh fading and a path loss exponent of 3.
3. Results and Discussion
The performance of the proposed algorithm was evaluated against an Exhaustive Search (ES) benchmark and a baseline scheduling method based on Shannon capacity.
As shown in the results, the proposed "Algorithm 1 with tuning" achieves near-optimal performance, closely matching the Exhaustive Search while maintaining significantly lower computational complexity ($O((KN_t)^3)$ vs. exponential).
Key observations include:
- Blocklength Sensitivity: Increasing the blocklength $n$ allows the system to accommodate more users, as more resources are available to meet the strict reliability requirements. For instance, at $n=1024$, the system schedules nearly double the users compared to $n=64$.
- Reliability Trade-off: A higher decoding error probability $\epsilon$ allows for more scheduled users, highlighting the fundamental trade-off between reliability and network capacity.
- Shannon Capacity Inaccuracy: The results confirm that scheduling based on Shannon capacity is overly optimistic, often failing to meet the actual reliability requirements of short-packet uRLLC.
4. Conclusion
This work demonstrates an effective optimization framework for maximizing user scheduling in ultra-dense uRLLC networks. By prioritizing set cardinality over sum rate and utilizing FBL theory, the proposed SCA-based algorithm provides a scalable solution for high-density 6G scenarios. Future work will focus on incorporating user fairness and extending the model to multi-cell environments with dynamic interference.
Summary updated