稳态大系统中关联预测法的改进

An Improvement of Interaction Predicting Method for Steady State Large Sized System

对于子系统的性能指标按一定顺序耦合的稳态大系统 ,采用两级递阶优化控制算法。利用关联预测法对子系统的关联进行解耦 ,利用下级子系统之间的横向信息传递对目标函数进行解耦。该方法是对关联预测法的改进 ,从而解决了一类不可分稳态优化问题。对算法的收敛性给出了严格证明 。

Two stage hierarchical optimization control algorithm is adopted for steady state large sized system with its subsystem performance indices coupling in a certain order. The interaction predicting method is used to decouple the subsystem interaction. The horizontal information transmission among the lower level subsystem to decouple the target functions, so that this method is an improvement to the interaction predicting method thereby solving the problem of a class of non separable steady state optimization, and the converqence of this algorithm is strictly proved and the simulation shows that algorithm posed is effective.