Balancing Accuracy and Data Reduction Using Record of Additional Data in Sequential Approximate Optimization

Product systems require optimization. In response to this, we have proposed a method of sequential approximate optimization using surrogate models. In this case, the amount of training data and solution accuracy are in conflict. In particular, the method of adding data to the vicinity of the optimal solution has not been sufficiently studied. Therefore, we will utilize the history of additional data and accuracy to achieve a good balance.