TSDM scheduling framework cracks throughput-AoI tradeoff in wireless nets

Low Age of Information (AoI) and high throughput are both critical for remote sensing and real-time monitoring over multichannel wireless networks. But jointly optimizing these metrics is notoriously difficult, especially when channels are heterogeneous and unreliable. In a new arXiv paper (arXiv:2608.05348), Lin Wang and I-Hong Hou introduce TSDM (Two-Stage Deficit Matching), a scheduling framework that tackles this challenge head-on. Instead of relying on simplistic averages, TSDM uses a second-order approach that characterizes each data flow by its mean and temporal variance—capturing both performance and jitter.

TSDM's first stage translates a high-level utility maximization objective into concrete target mean and temporal variance statistics for each node-channel pair. The second stage then applies a low-complexity Weighted Matching Deficit (WMD) rule to perform real-time channel assignment. The authors theoretically prove that TSDM achieves the desired mean and temporal variance for every flow. They also run extensive simulations on two open joint throughput-AoI optimization problems, demonstrating that TSDM significantly outperforms existing scheduling policies. This work offers a practical, provably effective method for balancing data freshness and network efficiency in real-world multichannel systems.