Abstract: Due to the deficiency of two coordination mode in Augmented Lagrangian Coordination (ALC) method that were a main problem in centralized ALC method to coordinate all the sub-problems leaded to the increase of couplings,and the sub-problems needed to be solved sequentially in the distributed ALC method,a Dominant Augmented Lagrangian Coordination (DALC) method was presented based on tradeoff between parallelism and efficiency.In DALC method,an existing sub-problem was selected as the master problem to coordinate other sub-problems.The principle of existing sub-problem selection and DALC's application conditions were discussed respectively,DALC's mathematical equivalence proof and convergence analysis were given in details.Experimental simulation results showed that the proposed method was effective and feasible,which provided a comprehensive reference for scholars to solve optimization problems such as complex mechanical design.

Key words: centralized augmented Lagrangian coordination method, distributed augmented Lagrangian coordination method, dominant augmented Lagrangian coordination method, complex system, optimization design

摘要： 鉴于增广拉格朗日协调方法中两种协调方式存在的不足,即集中式协调方式因引入主问题使耦合量增加而降低了协调执行效率,分布式协调方式因要求各子问题连续求解而降低了并行计算,提出兼顾效率与并行计算的主控式增广拉格朗日协调方法,并分别对主控问题的引入机制、适用条件和选择原则进行了讨论,同时给出了其模型数学等价性论证和收敛性分析。通过实例仿真结果表明了所提方法的有效性和可行性,为了解和使用该方法以解决复杂机械系统设计等优化问题提供了全面的参考。

关键词: 集中式增广拉格朗日协调方法, 分布式增广拉格朗日协调方法, 主控式增广拉格朗日协调方法, 复杂系统, 优化设计