基于一类SMSA策略的模型最优降阶  

Optimal reduction of models based on a class of SMSA strategy

作　　者：李令莱[1] 王凌[1] 郑大钟[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《控制与决策》2004年第8期947-950,953,共5页Control and Decision

基　　金：国家自然科学基金资助项目(60204008;60374060);国家973计划资助项目(2002CB312200);清华大学基础研究基金项目.

摘　　要：将一类SMSA混合策略应用于模型的最优降阶,提出了处理约束优化和解决模型降阶通常遇到的稳定性问题的有效方法.通过对典型例子的数值仿真表明了混合策略和约束处理方法的有效性,而且第3个例子的优化结果明显优于以往文献的结果.A class of SMSA hybrid strategy is applied to optimal reduction of models, and effective methods are also proposed to handle constraint and to solve stability problem that is usually encountered in reduced model. Numerical simulations based on benchmarks show the effectiveness of the hybrid strategy and constraint handling method, and the result obtained for the third benchmark is obviously better than that of literature.

关 键 词：模型降阶 模拟退火 单纯形法 约束处理 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]