Control system, control method, learning device, control device, learning method, and non-transitory computer-readable recording medium

A control method which is performed by a computer, the control method comprising:

• a step of acquiring a plurality of learning data sets of which each includes a combination of first data associated with a factor determining an operation of a subject device and second data associated with a command value for the subject device which is adapted to the factor indicated by the first data;
• a step of constructing a prediction model that outputs a value corresponding to the second data when the first data is input for each of the acquired plurality of learning data sets;
• a step of estimating a possible numerical value range of the command value from a distribution of the second data in the acquired plurality of learning data sets;
• a step of determining a second threshold value for the command value for the subject device on the basis of the estimated numerical value range such that a first acceptable range which is defined by a first threshold value preset for the command value for the subject device is extended;
• a step of acquiring input data associated with the factor in an operational phase;
• a step of acquiring an output value from the prediction model by inputting the acquired input data to the prediction model;
• a step of determining the command value for the subject device within a second acceptable range which is defined by the determined second threshold value on the basis of the acquired output value;
• and
• a step of controlling the operation of the subject device on the basis of the determined command value.

A learning device comprising:

• a learning data acquiring unit configured to acquire a plurality of learning data sets of which each includes a combination of first data associated with a factor determining an operation of a subject device and second data associated with a command value for the subject device which is adapted to the factor indicated by the first data;
• a learning processing unit configured to construct a prediction model that outputs a value corresponding to the second data when the first data is input for each of the acquired plurality of learning data sets;
• an estimation unit configured to estimate a possible numerical value range of the command value from a distribution of the second data in the acquired plurality of learning data sets;
• and
• a threshold value determining unit configured to determine a second threshold value for the command value for the subject device on the basis of the estimated numerical value range such that a first acceptable range which is defined by a first threshold value preset for the command value for the subject device is extended.

A learning method which is performed by a computer, the learning method comprising:

• a learning data acquiring step of acquiring a plurality of learning data sets of which each includes a combination of first data associated with a factor determining an operation of a subject device and second data associated with a command value for the subject device which is adapted to the factor indicated by the first data;
• a learning processing step of constructing a prediction model that outputs a value corresponding to the second data when the first data is input for each of the acquired plurality of learning data sets;
• an estimation step of estimating a possible numerical value range of the command value from a distribution of the second data in the acquired plurality of learning data sets;
• and
• a threshold value determining step of determining a second threshold value for the command value for the subject device on the basis of the estimated numerical value range such that a first acceptable range which is defined by a first threshold value preset for the command value for the subject device is extended.

A non-transitory computer-readable recording medium comprising a learning program causing a computer to perform:

• a learning data acquiring step of acquiring a plurality of learning data sets of which each includes a combination of first data associated with a factor determining an operation of a subject device and second data associated with a command value for the subject device which is adapted to the factor indicated by the first data;
• a learning processing step of constructing a prediction model that outputs a value corresponding to the second data when the first data is input for each of the acquired plurality of learning data sets;
• an estimation step of estimating a possible numerical value range of the command value from a distribution of the second data in the acquired plurality of learning data sets;
• and
• a threshold value determining step of determining a second threshold value for the command value for the subject device on the basis of the estimated numerical value range such that a first acceptable range which is defined by a first threshold value preset for the command value for the subject device is extended.