Abstract: Surrogate model is a substitute of the numerical analysis model with complex and time-consuming calculation, which can make up for the shortcomings of traditional engineering optimization design and numerical simulation analysis, such as long design cycle, high computing costs, and curse of dimensionality. Surrogate model technique has developed rapidly and achieved abundant outcomes in recent years. It has been gradually applied in aerospace, ship industry, vehicle, power generation and other fields. First, this paper gives a general overview of the surrogate model technique, and investigates the total and increment number of papers published on the surrogate model in the past thirty years. Second, the research status of surrogate model and related typical methods and theories are introduced in detail from three aspects, namely design of experiments, single-fidelity surrogate model and multi-fidelity surrogate model. Then applications of surrogate model in optimization design are investigated. The overall investigation of the surrogate-assisted multidisciplinary, surrogate-assisted multi-objective, and surrogate-assisted uncertainty optimization design techniques are analysed and summarized. Finally, the underlying problems existing in the surrogate model technique are summarized, and some suggestions are proposed for the future research.

Key words: surrogate model, design of experiments, multi-fidelity, sequential sampling, multi-objective optimization design, multidisciplinary optimization design