Control device, control method, and control program

A control method comprising:

• calculating an amount relating to a state of a control target detected by a sensor based on the state;
• detecting an abnormality predictor of the control target based on an amount obtained at a time when the amount relating to the state is input to a detection learning model;
• calculating a reward which is an amount obtained by evaluating pass or fail of an operation of the control target based on a detection result of the abnormality predictor;
• calculating a basic control amount serving as a reference to control the control target based on a control amount during an operation obtained at a time when data relating to the state during the operation is input to a calculation learning model;
• searching for a control amount at which the reward satisfies a predetermined condition based on the basic control amount, the data relating to the state during the operation, and the reward;
• outputting the searched control amount and a control parameter for execution of control based on the control amount;
• and
• generating a command value to the control target based on the searched control amount and the control parameter.

A control program on a non-transitory tangible computer readable medium that causes a computer to execute:

• calculating an amount relating to a state of a control target detected by a sensor based on the state;
• detecting an abnormality predictor of the control target based on an amount obtained at a time when the amount relating to the state is input to a detection learning model;
• calculating a reward which is an amount obtained by evaluating pass or fail of an operation of the control target based on a detection result of the abnormality predictor;
• calculating a basic control amount serving as a reference to control the control target based on a control amount during an operation obtained at a time when data relating to the state during the operation is input to a calculation learning model;
• searching for a control amount at which the reward satisfies a predetermined condition based on the basic control amount, the data relating to the state during the operation, and the reward, and outputting the searched control amount and a control parameter for execution of control based on the control amount;
• and
• generating a command value to the control target based on the searched control amount and the control parameter.