Deep reinforcement learning and parameter transfer based approach for the multi-objective agile earth observation satellite scheduling problem

Luona Wei, Yuning Chen, Ming Chen, Yingwu Chen

Abstract: Highlights•A deep reinforcement learning and parameter transfer based approach is proposed.•RLPT is the first attempt that applies deep reinforcement learning to a MO-AEOSSP.•RLPT shows better performance and computational efficiency against standard MOEAs.•RLPT is highly general and scalable on various-size instances.