Robust Deep Reinforcement Learning Scheduling via Weight Anchoring

Questions remain on the robustness of data-driven learning methods when crossing gap from simulation to reality. We utilize weight anchoring, a method known continual learning, cultivate and fixate desired behavior in Neural Networks. Weight anchoring may be used find solution problem that is nearby another problem. Thereby, can carried out optimal environments without neglecting or unlearning behavior. demonstrate this approach example mixed QoS-efficient discrete resource scheduling with infrequent priority messages. Results show provides performance comparable state art augmenting environment, alongside significantly increased steerability.