LEARNING PROGRAM, LEARNING METHOD, AND INFORMATION PROCESSING APPARATUS
To generate a learning model for estimating physical characteristics of a design target in a short period.SOLUTION: A processing unit 12 obtains a calculation value Va of physical characteristics of a design target calculated by a numerical calculation simulation based on a basic design condition fo...
Saved in:
Main Author | |
---|---|
Format | Patent |
Language | English Japanese |
Published |
28.10.2021
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | To generate a learning model for estimating physical characteristics of a design target in a short period.SOLUTION: A processing unit 12 obtains a calculation value Va of physical characteristics of a design target calculated by a numerical calculation simulation based on a basic design condition for the design target, and generates a plurality of design change conditions in which the basic design condition is changed. The processing unit 12 obtains calculation values Vb1 to Vbn of the physical characteristics calculated by the numerical calculation simulation based on each of the plurality of design change conditions, and calculates evaluation values representing a magnitude of change of each of the calculation values Vb1 to Vbn to the calculation value Va. Then, the processing unit 12 generates a learning model 11b by learning using training data 11a in which information representing each of the plurality of design change condition is set to an explanatory variable and the evaluation value corresponding to each of the plurality of design change conditions is set to an objective variable.SELECTED DRAWING: Figure 1
【課題】設計対象の物理的特性を推定する学習モデルを短期間で生成する。【解決手段】処理部12は、設計対象の基本設計条件に基づく数値計算シミュレーションにより計算される設計対象の物理的特性の計算値Vaを取得し、基本設計条件を変更した複数の設計変更条件を生成する。そして、処理部12は、複数の設計変更条件のそれぞれに基づく数値計算シミュレーションにより計算される物理的特性の計算値Vb1〜Vbnを取得し、計算値Vaに対する計算値Vb1〜Vbnのそれぞれの変化の大きさを表す評価値を計算する。そして、処理部12は、複数の設計変更条件のそれぞれを表す情報を説明変数、複数の設計変更条件のそれぞれに対応する評価値を目的変数に設定した訓練データ11aを用いた学習により、学習モデル11bを生成する。【選択図】図1 |
---|---|
Bibliography: | Application Number: JP20200072211 |