Intelligent Interactive Physical Programming And It’s Application In MDO

With the rapid development of high technology field, there are many differentdisciplines which we should consider when designing a complex product．Tranditionally,we devide the product into many subsystems based on the disciplines．Differentdepartments will accomplish the design of different subsystems．Because when thedesigners try to find the best solution in their discipline, they often ignore thecoinfluence of other disciplines．In that way, they can only get a local-best result instead of the real best result that we desire for with long cycle and high cost. Inlarge-scale complex engineering system, the traditional mode is hard to meet people’sdemands and needs．In this background, multidisciplinary design optimization (MOO)methods emerged.The main idea of MDO is to consider the coupling effects among differentdisciplinary and reach the optimized solution through interactions and collaborationsamong all. But there are also problems in MDO such as low computing efficiency anddifficulties to obtain the optimal solution features with preferences of designers.To resolve the above-mentioned problems, in this thesis our work are as follows:(1) Establish and realize the intelligent decision-making support model usingFuzzy Neural Networks (FNN), study and improve the BP neural network. ThePCA-GA-BP algorithem, Adaptive Construction GA-BP (AC-GA-BP) algorithm andAC-PCA-GA-BP algorithem are presented to construct the structure of the hidden layerin the BP neural network and to improve the BP algorithm in the thesis.Intelligent Interacitve Physical Programming (IIPP), based on IPP and intelligentdecision-making support method is established. And the simulation results prove thefeasibility of this approach.(2) Integrated with the Bi-Level Integrated System Synthesis (BLISS), IntelligentInteracitve Physical Programming (IIPP) is utilized in MDO. A new method is proposedto solve the multidisciplinary and multi-objectives design optimization. Finally, thepractical application shows the feasibility of the new method.