Research on digital twin based key technologies for multi-AGV scheduling and cooperation in complex systems

This project investigates advanced technologies for multi-tasking and multi-operation collaborative scheduling of AGV fleets in complex dynamic scenarios. It is aimed at addressing key issues in the current AGV industrial applications, including poor compatibility of multiple brands and models, poor global optimality, poor real-time dynamic capabilities of algorithms, and insufficient versatility in multiple scenarios. We aim to mainly focus on hyper-heuristic scheduling technology based on hierarchical multi-granularity simulation and BDI multi-agents, trained through data-driven dynamic weight multi-objective deep reinforcement learning optimization.