Learning Resource Scheduling with High Priority Users using Deep Deterministic Policy Gradients

Abstract

Advances in mobile communication capabilities open the door for closer integration of pre-hospital and inhospital care processes. For example, medical specialists can be enabled to guide on-site paramedics and can, in turn, be supplied with live vitals or visuals. Consolidating such performance-critical applications with the highly complex workings of mobile communications requires solutions both reliable and efficient, yet easy to integrate with existing systems. This paper explores the application of Deep Deterministic Policy Gradient (DDPG) methods for learning a communications resource scheduling algorithm with special regards to priority users. Unlike the popular Deep-Q-Network methods, the DDPG is able to produce continuous-valued output. With light post-processing, the resulting scheduler is able to achieve high performance on a flexible sum-utility goal.

Robust Deep Reinforcement Learning Scheduling via Weight Anchoring

“…We implement a DDPG-based scheduler, as used in our previous work [2], to learn to output allocations A t that approximately optimize the sum reward r t in (3). The scheduler uses a pre-processor, two neural networks (actor and critic), a memory module, an exploration module, and a learning module.…”

“…In the field of communication systems, deep Reinforcement Learning (RL) has stirred interest with its ability to learn approximately optimal strategies with limited model assumptions. This is an auspicious promise for challenges that are complex to model or to solve in real time, and early research has shown that deep RL strategies are indeed applicable to problems such as intelligent resource scheduling [1], [2]. Ultra-reliability, low latency, and heterogeneous QoSconstraints are cornerstones of future communication systems, particularly in fields such as medical communications, and flexible and fast learned schedulers may aid in achieving required performance goals.…”

Machine Learning and Deep Learning powered satellite communications: Enabling technologies, applications, open challenges, and future research directions

“…Optimal precoding requires accurate CSI and robust precoding can be implemented by using imperfect CSI. 105 The authors of Gracla et al 127 suggest a soft actor-critic deep reinforcement learning model. This model is tested against minimum MSE precoding and orthogonal multiple access pre-coding.…”

“…The authors of Gracla et al 127 suggest a soft actor‐critic deep reinforcement learning model. This model is tested against minimum MSE pre‐coding and orthogonal multiple access pre‐coding.…”

Deep Reinforcement Model Selection for Communications Resource Allocation in On-Site Medical Care

“…We define our first task, that we wish to learn to satisfaction and then prevent unlearning, as the handling of priority messages with high reliability. Subsequently, we then integrate two methods from multi-task learning with a vanilla deep RL resource scheduler which we used in [14]. We show that the two methods, 1) Elastic Weight Consolidation (EWC) [13], presented in our work [15].…”