CONTROLLER AND MACHINE LEARNING DEVICE

A controller is provided with a repeated control unit configured to calculate a position compensation value, based on a position command for each control period of a motor, a position deviation, which is the difference between the position command and a position of the motor, and a parameter for repeated control, and a machine learning device configured to predict the position compensation value calculated by the repeated control unit, and the machine learning device constructs a learning model so as to minimize an objective function based on the position command, the position deviation, and the position compensation value.