Uncertainty Multi-objective Optimization Of Frame Based On Approximate Model

Uncertainty widely exists in practical engineering problems. Considering the impact of the uncertainties on the target response has an important significance in engineering problems. There are mainly four models to deal with uncertainty,which are probability model,fuzzy model and convex model or uncertainty model.A great amount of sampling information on the uncertainty is required to construct the precise probability distributions or fuzzy membership functions in the probility optimization and fuzzy optimization. Unfortunately,it often seems very difficult or sometimes expensive to obtain sufficient uncertainty information in applicability. Interval optimization has a better economy and convenience in which the variation bounds of the uncertain variables are only required.Frame is an important assembly bearing loads of an automobile. All kinds of loads will pass to it.So the performance of frame structure affects whether the automobile design is successful or not. The frame material in the manufacturing process, inevitably there are material properties, boundary conditions, initial conditions, measuring the deviation error or uncertainty. Although these error or uncertainty smaller value, in most cases, but are coupled together may make the system response to a large devia-tion. Therefore consider the frame parameters uncertainties affect the design of the frame has important theoretical and practical significance.In this paper,considering under the influence of uncertainties in the frame parameters, Considering the uncertainty of frame pa-rameters, the frme is carried out using multi-objective optimizat-ion in this paper. Build an approximate model between design variables, uncertain variables and objective function. The uncerta-n material parameters of frame are treated as intervals.Uncertainty multi-objective optimization model of the frame is established,and it is transformed into a deterministic multi-objective optimization model based on the order relation of interval number. On the basis of the approximate model,the nesting optimization problem is solved using non-dominate sorting genetic algorithm (NSGA-II) which is improved by adding elitist preserve strategy and removing duplicates individuals algorithm and IP-GA. Taking into account the low efficiency of double nested,nesting optimization problem of uncertainty multi-optimization is transformed into a singlelayer optimization based on the interval anlysis method. Similarly NSGA-II as outer layer of the multi-objective optimization algori-hm, again based on the approximate model, solving the multi-objective optimization problem of the uncertainty of the frame, and this method can greatly improve the solution efficiency. Com-pare the results obtained by the two methods, which can show sup-eriority of uncertain optimization. Lastly, using interval analysis method to get a set of solution set in finite element analysis, to verify the optimization results are effective and feasible.