Data management of scientific applications in a reinforcement learning-based hierarchical storage system

Tianru Zhang, Ankit Gupta, María Andreína Francisco Rodríguez, Ola Spjuth, Andreas Hellander, Salman Toor

Abstract: Highlights•The challenges of data management in scientific applications are discussed.•Adaptive hierarchical storage management framework with reinforcement learning.•Experiments are carried out based on four real scientific datasets.•We have compared proposed framework with five widely-used methods.•Results highlight that our framework is superior in reducing data processing time.