围绕由多构件组成的复杂结构多学科优化与可靠性设计工程背景，本项目从函数逼近和工程设计角度，深入研究多层次、多模型协同优化与可靠性设计的分布式协同代理模型方法，解决处理设计变量多、非线性程度高和多层循环嵌套迭代所带来的计算精度和效率问题，实现复杂结构多构件、多学科协同优化与可靠性设计。主要内容包括：多层复合函数逼近的分布式协同代理模型研究、基于分布式协同代理模型的多层次多模型协同优化方法研究和基于分布式协同代理模型的多层次多模型协同可靠性设计方法研究，并以复合函数概率统计分析、涡轮叶盘多学科优化和涡轮叶尖间隙可靠性设计为例，加以验证。本项目为解决多层次、多模型协同优化与可靠性设计这一类问题提供技术手段。研究成果对概率数理统计理论和机械可靠性设计方法的丰富和发展具有重要的学术价值，在航空、航天、机械各工程领域的复杂结构（系统）设计中有现实的应用前景。

Around the engineering background of the multi-discipline optimization and reliability design of complex structure comprising many components, this project deeply investigates distributed collaborative surrogate model methodology for multi-layer multi-model collaborative optimization and reliability design from the perspectives of function approximation and engineering design, to address the problem of computational accuracy and efficiency resulting from many design variables, high nonlinearity and multilayer nested loop iterations, and carry out the multi-component multi-discipline collaborative optimization and reliability design of complex structures. The main contents include the studies on the distributed collaborative surrogate model of multi-layer compound function approximation, multi-layer multi-model collaborative optimization approach based on distributed collaborative surrogate model and multi-layer multi-model collaborative reliability design approach based on distributed collaborative surrogate model. The developed methodologies are validated by compound function probabilistic statistical analysis, turbine blisk multidiscipline optimization and turbine blade-tip clearance reliability design. This study offers a technological mean for the type of multi-layer multi-model collaborative optimization and reliability design. The efforts of this project hold vital academic significances in enriching and developing the theory and method of probability and mathematical statistics, which has the real application prospect in complex structure(system) design in the engineering fields of aviation, aerospace, machine and so forth.