Self-Learning Metamodels for Optimization

The paper discusses strategies for metamodels in optimization problems involving expensive high-delity analysis tools. A strategy to assess the accuracy of the metamodels and to increase their completeness and delity during the optimization process is suggested and evaluated.