Composite Dispatching Cost-Based Time-Sensitive Frame Aggregation Scheduling for Multiqueue Wireless Communications

Frame aggregation (FA) significantly enhances throughput by frame header reduction and payload compression. However, FA affects the performance of latency and deadline adherence due to aggregation delays. In this paper, we investigate time-sensitive FA scheduling to ensure low latency and low delay violation probability while maintaining high throughput. We formulate the complex FA scheduling problem by constraint programming (CP), which considers variable FA sizes and queue availability in FA scheduling and yields near-optimal solutions. To address the NP-hard problem imposed by CP, we propose a composite dispatching cost - genetic algorithm (CDC-GA) for FA scheduling. According to our theoretical analysis, the proposed CDC encourages more aggressive aggregation when frame header overhead is high, thereby optimizing for both throughput and time-centric performance. The GA then refines the initial schedule using a proposed ternary chromosome encoding, which comprehensively captures all necessary scheduling decisions. The proposed CDC-GA approach accommodates various FA structures (e.g., fixed/variable length, with/without data compression), and outperforms the existing approaches, both conventional and learning-based, by adeptly handling variable FA sizes and integrating the impact of FA on time-sensitive performance. Simulation results show that the proposed CDC-GA significantly outperforms the existing approaches, improving throughput by 77% and reducing average latency by 36%.