The high fidelity numerical simulation codes used in aerodynamic design and optimization applications show a number of drawbacks. Of these, the high computational cost, memory and time demands associated with complicated designs. Metamodels or surrogates provide a much cheaper alternative for these codes. The use of spikes in the design of hypersonic vehicles yields a considerable reduction in drag and aerodynamic heating effects. In this work, surrogates are used in the context of design optimization of a spiked blunt body in hypersonic flow conditions. Four different surrogate models including quadratic response surface and kriging were constructed based on the values of drag and heating responses. The evolutionary genetic algorithm is applied to find the optimum design based on each of the surrogates. The structure of the various models was investigated and the main differences were addressed. For the cases investigated in this work, kriging surrogate based on exponential correlation produced a relatively better prediction of new points in the
design space and, consequently, better optimized design.