| 摘要 |
PROBLEM TO BE SOLVED: To apply the past analysis result as an analysis result with regard to the result of a new calculation in a control parameter adaption method for a device to be controlled, and thereby achieve the accuracy and efficiency of an optimization calculation at a still higher level.SOLUTION: The present invention provides a control parameter adaption method for adding experiment data (additional data) about a device to be controlled to existing data (initial data) about a similar device and updating an approximate equation. When there is additional data in which deviation from an existing approximate equation is large, "active learning" is performed that deletes initial data in which deviation from the approximate equation updated by the additional data is large, and the approximate equation is updated on the basis of the remaining initial data and the additional data.SELECTED DRAWING: Figure 3 |
