Input parameter value set identifying apparatus and method

An input parameter value set identifying method, comprising:

• calculating, for each of input variable value sets, which are set in advance, an indicator value that is a standard deviation of actually measured output variable values from data stored in a first storage device, and storing the calculated indicator value into a second storage device, wherein said first storage device stores, for each of said input variable value sets, the actually measured output variable values;
• repeating a processing to cause a simulator for computing a calculated output variable value that is a calculated value of an output variable from a certain input parameter value set and a certain input variable value set to compute a calculated output variable value for each combination of one of candidate input parameter value sets and one of said input variable value sets, and a processing to calculate, for each of said candidate input parameter value sets, a first entire error that is a result obtained by taking into consideration, with respect to all of the input variable value sets, first partial errors obtained respectively by evaluating, by using the indicator value for a corresponding input variable value set, which is stored in the second storage device, a difference between the calculated output variable value and the actually measured output variable value for the corresponding input variable value set, which is stored in said first storage device, or a second entire error that is a result obtained by taking into consideration, with respect to all of the input variable value sets, second partial errors, each of which is said difference, while weighting by using the indicator value for the corresponding input variable value set;
• identifying the candidate input parameter value set making the first or second entire error minimum, and storing the identified candidate input parameter value set into an output data storage device.

An input parameter value set identifying apparatus, comprising:

• a first storage device storing, for each of input variable value sets, actually measured output variable values;
• an indicator calculation unit to calculate, for each of said input variable value sets, which are set in advance, an indicator value that is a standard deviation of the actually measured output variable values from data stored in said first storage device, and store the calculated indicator value into a second storage device;
• and
• a controller to repeat a processing to cause a simulator for computing a calculated output variable value that is a calculated value of an output variable from a certain input parameter value set and a certain input variable value set to compute a calculated output variable value for each combination of one of candidate input parameter value sets and one of said input variable value sets, and a processing to calculate, for each of said candidate input parameter value sets, a first entire error that is a result obtained by taking into consideration, with respect to all of the input variable value sets, first partial errors obtained respectively by evaluating, by using the indicator value for a corresponding input variable value set, which is stored in the second storage device, a difference between the calculated output variable value and the actually measured output variable value for the corresponding input variable value set, which is stored in said first storage device, or a second entire error that is a result obtained by taking into consideration, with respect to all of the input variable value sets, second partial errors, each of which is said difference, while weighting by using the indicator value for the corresponding input variable value set;
• and
• wherein said controller identifies the candidate input parameter value set making the first or second entire error minimum, and stores the identified candidate input parameter value set into an output data storage device.

An input parameter value set identifying apparatus, comprising:

• a memory configured to store, for each of input variable value sets, actually measured output variable values;
• and
• a processor configured to execute a procedure, the procedure comprising:
• calculating, for each of said input variable value sets, which are set in advance, an indicator value that is a standard deviation of the actually measured output variable values from data stored in said memory, and store the calculated indicator value into the memory;
• and
• repeating a processing to cause a simulator for computing a calculated output variable value that is a calculated value of an output variable from a certain input parameter value set and a certain input variable value set to compute a calculated output variable value for each combination of one of candidate input parameter value sets and one of said input variable value sets, and a processing to calculate, for each of said candidate input parameter value sets, a first entire error that is a result obtained by taking into consideration, with respect to all of the input variable value sets, first partial errors obtained respectively by evaluating, by using the indicator value for a corresponding input variable value set, which is stored in the memory, a difference between the calculated output variable value and the actually measured output variable value for the corresponding input variable value set, which is stored in said memory, or a second entire error that is a result obtained by taking into consideration, with respect to all of the input variable value sets, second partial errors, each of which is said difference, while weighting by using the indicator value for the corresponding input variable value set;
• and
• identifying the candidate input parameter value set making the first or second entire error minimum, and stores the identified candidate input parameter value set into the memory.