Research Article
A Pinning Actor-Critic Structure-Based Algorithm for Sizing Complex-Shaped Depth Profiles in MFL Inspection with High Degree of Freedom
(i) | Initialize actor network , critic network , target network and , replay buffer | (ii) | For episode = 1, M do | (iii) | Initialize pinning subdefects , interpolate to have the full depth profile | (iv) | Get initial observation state from reference signal and depth of sub-defects | (v) | For t = 1, T do | (vi) | Generate an action from the output of actor network and exploration noise process | (vii) | Execute action , obtain new depth of pinning sub-defects | (viii) | Interpolate to get the full depth profile within the ROI, calculate reward and new state | (ix) | Store in | (x) | If capacity of replay buffer is full then | (xi) | Randomly sample piece of data from | (xii) | Update the critic network and actor network with (5) and (3) | (xiii) | Update the target networks: | (xiv) | | (xv) | end if | (xvi) | If error between each reference subdefect and reconstructed subdefect is less than , then | (xvii) | break | (xviii) | end if | (xix) | end for | (xx) | end for |
|