Path Planning of Unmanned Helicopter in Complex Environment Based on Heuristic Deep Q-Network

Algorithm 1

DQN algorithm.

Algorithm: DQN algorithm

Initialization: initialize training network parameter and target network parameter , .

Iterative process:

Repeat (for each episode)

   Initialization state

    Repeat (for each step)

      Select action based on the policy

      Perform action to get reward and next state

      Store transition in the experience memory

      Sample random mini batch from the experience memory

      Loss function is obtained

      Updating network parameters

    End Repeat (is the terminated state)

End Repeat (End of the training)