Precise and Adaptable: Leveraging Deep Reinforcement Learning for GAP-based Multipath Scheduler | IEEE Conference Publication | IEEE Xplore