Wing-body optimization based on multifidelity surrogate model

This paper focuses upon the efficient surrogate model algorithm for expensive simulation-based design optimization problems. Co-kriging method is used to develop a multi-fidelity surrogate model using two independent datasets. To achieve this objective, wing-body problem is taken as an example of application for highdimensional complex design problem. In addition, a simple sampling analysis is used to demonstrate the characteristics of co-kriging multi-fidelity surrogate model based on the defined criteria. A drag reduction optimization is carried on using genetic algorithm based on the co-kriging surrogate model. The results are compared with kriging model based optimization. It is shown that the integration of multi-fidelity surrogate model into evolution algorithm provides an efficient framework for design and analysis of expensive simulationbased design optimization problems.