Abstract: The automobile lightweight is one of the development directions of automobile industry, and also is an important symbol of technical progress and advanced extent of an automobile manufacturer and a country. The structural design optimization of the automobile is an effective way of realizing the automobile lightweight. The sample data design is carried out by adopting Latin-square experiment design method, moreover, in order to improve computing efficiency, a response surface approximate model based on moving least square fitting is introduced into the complex system of optimization design of complete vehicle frontal crash, and the sequential response surface optimization method is used to optimize the approximate model, thus avoiding the disadvantages of large amount of circulation in the traditional optimization method and slow convergence or even no convergence in the optimization of nonlinear crash system. While making the vehicle meet the requirements of Regulation CMVDR 294, it also makes the plate thickness of some components of automobile body achieve rational distribution, thus reducing the vehicle weight to some extent.

Key words: Latin-square experiment design, Lightweight, Moving least square, Sequential response surface