A Multi-objective Optimization Method Based On Adaptive Approximation Model And Its Application In Vehicle Engineering

In the rapid development of China's new century,the optimization technology has been widely used,and the optimization problems in engineering often have multiple goals that need to be optimized,and the contradictions between the goals are difficult to unify,so the difficulty of solving the problem is increased.The traditional single objective optimization method can not meet problems,and the multi-objective optimization method based on intelligent algorithm came into being.The intelligent algorithm can search the optimization result without considering the specific characteristics of the problem,and is suitable for dealing with multi-objective problems in engineering.It requires more iteration calculations due to the influence of model data size and convergence speed.Many objectives in engineering cannot be explicitly expressed,so a multi-objective optimization method based on approximate model is proposed.Through the application of the approximate model,the number of calling the real model can be greatly reduced,and the efficiency of solving the problem can be accelerated.However,there are many kinds of approximate models,and the construction process is also very complicated.Some approximate models are affected by the parameters and cannot be solved accurately.The adaptive approximate model developed in recent years can improve the accuracy of the model through continuous optimization.Therefore,this paper studies the multi-objective optimization method based on adaptive approximate model.The research will be carried out from the following aspects:Firstly,a low cost is proposed for the shape parameter problem of radial basis function in approximate model construction.The high-precision calculation method combines the local encryption method and the micro multi-objective genetic algorithm to obtain the non-inferior solution of the multi-objective optimization problems.Secondly,a model error prediction method based on the adaptive approximate model is proposed to further reduce the non-inferior solution.In the end,the multi-objective optimization design method based on adaptive approximate model is applied to the safety design of the cab structure of heavy trucks.The specific research contents of this paper are as follows:1.A multi-objective optimization algorithm based on approximate model is proposed.In order to solve the problem of approximate model construction in optimization problems,the inverse parametric method is proposed to solve it accurately.On the basis of not calling the real model without adding additional sampling points,the sample point information is fully utilized to obtain the optimal shape parameters in the current sample space.Finally,a comparative numerical example shows the advantage of this method over the traditional formula method.An uncertain multi-objective optimization example shows that the method is equally applicable to uncertain problems,and then through the I-beam example and gear structure.Optimize the design example to prove the method.2.A multi-objective optimization strategy based on approximate model is proposed.Firstly,the approximate model of the objective function is constructed by using the radial basis function approximation model method.The local encryption method is used to encrypt the algorithm at the initial stage.This chapter proposes a model error prediction method based on the error characteristics of non-inferior solutions.The final step can be predicted,and the predicted value is substituted for the optimal solution output of the approximate model.Finally,the effectiveness of the method is illustrated by the I-beam example and the gear structure optimization.This method is especially suitable for the case where the error distribution is dense.3.The multi-objective optimization method based on radial basis agent model is combined with modem auxiliary design software to apply to the cab design of Anhui Hualing truck.The cab numerical model is established through the preparation of pre-modeling,drawing grid,etc.,and multi-objective optimization problems are established for quality and member living space.Taking the key plate thickness as the design variable,the overall quality and compression distance of the cab are taken as the optimization target,and the agent model is constructed and optimized.Firstly,the approximate model is constructed by LHD and inverse parameter method,then iterative optimized by local cryptography and ?MOGA,and then the model error prediction method is used to predict the non-inferior solution to obtain the final design size.Finally,12 sets of optimal solutions are listed and the choices based on different design preferences are discussed.