This paper describes the use of Reinforcement Learning to the computation of optimal trajectories and anti-swing control of a ship unloader. The unloading cycle is divided into six phases and an optimization problem is defined for each of them. A TD(0) algorithm together with a multilayer perceptron neural network as a value function approximator is used in the optimization. The results obtained are compared to Optimal Control results.
Reinforcement Learning; Optimal Control; Anti-Swing Control; Ship Unloaders; Neural Networks
