Some Improvement Strategy Research Based On The Collaborative Optimization Algorithm

Human needs has always been one of the main power of social development,whether from the angle of design and manufacture,there is no engineering not to meet the needs of human beings as the main objective.Nowadays,the demand of human beings is no longer limited to functional satisfaction,and gradually to the diversification of the demand,which makes a separate product integrated more and more disciplinary knowledge.The integration of multidisciplinary knowledge greatly increases the difficulty of product design and optimization,and multidisciplinary design optimization is just an effective way to solve this problem.This paper introduces the research background and the present situation of the research on the multidisciplinary design optimization,and simply describes the development history and the development process of some representative algorithm,mainly introduces the cooperative optimization algorithm in multidisciplinary design optimization,describes its basic idea and framework,and also put forward three improved algorithm for the shortcomings that the optimized results are sensitive to initial points and the optimization process to fall into local optimum easily.The first improved algorithm integrate the disciplinary optimal design points into the disciplinary optimizing process.The improved algorithm avoids disciplinary optimizing process only use system level optimal design point as the reference points of disciplinary optimizing process,which increases feasible regions of disciplinary optimizing process,reduces the optimization result dependence of the initial points and also reduces the possibility to get local optimum.we call the improved algorithm as static disciplinary optimal design points collaborative optimization in this paper.The second improved algorithm depends on the static disciplinary optimal design points collaborative optimization,which makes the disciplinary reference points dynamic and use the last disciplinary optimal design points as the reference points of disciplinary optimizing process.The improved algorithm has all advantages of static disciplinary optimal design points collaborative optimization and has better stability and convergence.we call the second improved algorithm as dynamic disciplinary optimal design points collaborative optimization in this paper.The third improved algorithm is based on the response surface collaborative optimization algorithm,which integrate confidence regions into the constructing process of response surface to improve the efficiency to select test design points.We call the third improved algorithm as modified confidence regions response surfacecollaborative algorithm in this paper.In summary,we analyze and compare the three kinds of improved algorithms,state the characteristics,the differences and the application environment.