Storage medium, information processing method, and information processing apparatus

An information processing method for a computer to execute a process comprising:

• acquiring a first model trained using a first training data group to predict, from a combination of (i) constituent values of a target object and (ii) an environmental value in an experiment on the target object, a mean value of a characteristic value of the target object and a deviation value of the characteristic value, the first training data group including first training data pieces, each piece of the first training data pieces including a characteristic value of a respective first target object and a respective first combination of (i) constituent values of the respective first target object and (ii) a first environmental value in an experiment on the respective first target object;
• acquiring a second model trained using a second training data group to predict an allowable condition for the experiment on the target object from the constituent values of the target object, the second training data group including second training pieces, each of the second training pieces including (i) constituent values of a respective second target and (ii) an allowable condition for an experiment on the respective second target;
• and
• generating a solution set of the combination of the constituent values and the environmental value for the target object by performing multi-objective optimization using (i) a penalty term based on the allowable condition identified by the acquired second model, (ii) a first objective function for searching for a value of a first objective variable based on the mean value identified by the acquired first model, and (iii) a second objective function for searching for a value of a second objective variable based on the deviation value identified by the acquired first model.

An information processing apparatus comprising:

• one or more memories;
• and
• one or more processors coupled to the one or more memories and the one or more processors configured to:
• acquire a first model trained using a first training data group to predict, from a combination of (i) constituent values of a target object and an environmental value in an experiment on the target object, a mean value of a characteristic value of the target object and a deviation value of the characteristic value, the first training data group including first training data pieces, each piece of the first training data pieces including a characteristic value of a respective first target object and a respective first combination of (i) constituent values of the respective first target object and (ii) a first environmental value in an experiment on the respective first target object;
• acquire a second model trained using a second training data group to predict an allowable condition for the experiment on the target object from the constituent values of the target object, the second training data group including second training pieces, each of the second training pieces including (i) constituent values of a respective second target and (ii) an allowable condition for an experiment on the respective second target;
• and
• generate a solution set of the combination of the constituent values and the environmental value for the target object by performing multi-objective optimization using (i) a penalty term based on the allowable condition identified by the acquired second model, (ii) a first objective function for searching for a value of a first objective variable based on the mean value identified by the acquired first model, and (iii) a second objective function for searching for a value of a second objective variable based on the deviation value identified by the acquired first model.