中文摘要：

针对在模块化平台中筛选共享变量的问题，受协同优化算法使用一致性约束函数对各子优化问题进行一致性统筹的启发，提出一种基于一致性约束的优化算法．在系统级优化中使用非支配排序的遗传算法NSGA-Ⅱ对该多目标优化问题进行求解后得到Pareto解集，利用模糊聚类算法对解集中的每组解进行综合性能的评价并选优，根据最终筛选出的最优解即可实现共享变量的筛选．相比以往常用的基于经验或灵敏度的方法，该方法更严格地在子学科优化中以车身性能为目标函数，在系统级优化中进行共享度的优化，并且可根据系统级优化结果筛选出局部共享变量．以SUV、两厢掀背车和三厢轿车为算例，使用该方法有效地筛选出3款车型的全局共享变量、局部共享变量和非共享变量，对该方法的可行性和有效性进行了验证．

英文摘要：

For sharing modules selection in the modular platform design, an algorithm based on con- straints was proposed, which was inspired by Collaborative Optimization using the constraints to coordi- nate the System Level Optimization and Sub-system Level Optimization. In this algorithm, NSGA-Ⅱ was used in the System level to solve the Multidisciplinary Design Optimization problem, and a Pareto set was gotten. By using the fuzzy set theory to evaluate each equation in the Pareto set, the optimum solution was easily picked out, and the sharing modules selection was realized in this way. At last, this method was verified by using an application example constituted by involving a SUV, a hatchback, and a sedan.