The Multi-objective Reliability-based Design Optimization Based On Probability And Ellipsoidal Convex Hybrid Model

Reliability based design optimization problem exists widely in engineering problems.It not only meets the requirement of reliability,but also makes the design objective optimal.Therefore,reliability-based design optimization is widely used.The traditional reliability optimization design is usually a single objective problem,but in practical engineering,most of them are multi-objective optimization problem.The traditional reliability problem is generally a probabilistic reliability problem,and in practical engineering,a single probability variable can not meet the design requirements.Errors such as production,manufacturing,measurement,installation,etc.make the variables fluctuate within a certain range,and then use the non-probabilistic convex model to describe the variables.In view of the large amount of sample information needed by probabilistic models,non probabilistic convex models show their advantages.In the non-probabilistic model,it does not require too much sample information,but only the basic interval of variables.The non-probabilistic convex models which are commonly used are interval model and ellipsoid model.The interval model describes the interval information of independent variables,and the ellipsoid model can not only describe the interval information variables,but also reflects the correlation between variables.Therefore,the multi-objective reliability-based design optimization based on probability and ellipsoidal convex hybrid model is studied in this paper.The main research contents and achievements are as follows:1.A nested optimal decoupling strategy is proposed,and the three nested optimization problems are converted into two nested optimization problems,which greatly improves the efficiency of the solution.The micro multi-objective genetic algorithm and intergeneration projection genetic algorithm respectively as optimization solver inside and outside layer,to obtain the global optimal solution.2.A hybrid reliability multi-objective optimization design method based on local encryption approximation model is proposed.The radial basis function approximation model is used to construct the approximate model of the constraint function and the objective function respectively,use the approximate model instead of the real model.According to the result of the solution,local encryption is carried out for the less accurate part,so that the optimization results with higher precision are obtained,and the optimization efficiency and accuracy are improved.3.Based on Visual Basic 6.0 programming language,The multi-objective reliability-based design optimization based on probability and ellipsoidal convex hybrid model platform is established.A dynamic model of the airbag is established,and it is applied to the optimization design of airbag for airdrop equipment.