A Period Training Method for Heterogeneous UUV Dynamic Task Allocation

In the dynamic task allocation of unmanned underwater vehicles (UUVs), schemes UUVs need to be quickly reallocated respond emergencies. The most common heuristic method uses predesigned optimization rules iteratively obtain a solution, which is time-consuming. To assign tasks heterogeneous UUVs, we propose novel algorithm based on multi-agent reinforcement learning (MARL) and period training (PTM). (PTM) used optimize parameters MARL models in different environments, improving algorithm’s robustness. simulation results show that proposed methods can effectively allocate within few seconds reallocate real time deal with