Path Planning Algorithm for Unmanned Surface Vessel Based on Multiobjective Reinforcement Learning

Table 2

Training parameters for reinforcement learning.

Scene	Learning rate	Batch_size	Discount factor	FC structure
Dynamic obstacle avoidance	0.0005	256	0.99	[64, 64, 32]
Goal approaching	0.001	256	0.99	[32, 32, 16]
Main scene	0.0005	256	0.99	[32, 32, 16]