Utilizes a comprehensive system identification model developed through behavioral cloning to incorporate system dynamic properties.\ Introduces a demand-driven hierarchical integrated planning-scheduling control framework for FSMS based on the model. \ The two interconnected feedback control loops, an outer planner loop and an inner scheduler loop, can synergistically generate integrated planning-scheduling policies.\ Presents a novel P-MADDPG algorithm that leverages system properties to improve scheduling.
Lab Contributors: Chen, Kshitij, Muhammad