Abstract: In order to find optimum aerodynamic shapes with reasonable computational costs to yield a substantial improvement, this paper presents a surrogate\|based multi\|objective design optimization approach for conceptual design of a lifting\|body configuration using geometric class function/shape function transformation technique and global optimization strategy. The class function/shape function can adjust a small amount of dimensional variables or class function/shape function variables to develop the various design constructed shapes. Optimal Latin hypercube sampling is used as design of experiments strategy, radial basis functions are selected as approximation method and the surrogate model is modified by sequential sampling which improves goodness of fitting. Surrogate model of aerodynamic analysis for a brief representation of a lifting\|body configuration is built at Mach 6 with an angle of attack of 20°. The multi\|objective evolutionary algorithm NSGA\|Ⅱ which produces a set of Pareto solutions is employed as optimization strategy to obtain trade\|offs between maximization of lift\|to\|drag\|ratio and volumetric efficiency. Relative errors between objective responses of the surrogate model and real function are within 6.1% and it demonstrates that surrogate\|based multi\|objective design optimization not only can maintain good computational efficiency, but also increase the global search capability of the optimization, meet precision requirements of conceptual design and very useful to provide guidance on how to change aerodynamic shape and quantify tradeoffs among conflicting objectives effectively.

Key words: Aircraft conceptual design, Lifting-body configuration, Class function-shape function, SurroGate model, Multi-objective optimization, Aerodynamic shape optimization