摘要
分析了时变目标函数及其最优解的时变性能，提出了跟踪时变目标函数最优解的一阶和二阶方法，其中二阶方法能使时变目标函数最优解的跟踪以指数速率收敛。
The variation of a time dependent objective solution is analyzed. A first order technique and a second order technique for tracking the optimum are proposed. The second order technique ensures the tracking to converge exponentially. Applying time dependent optimization, the continuous linear system without time delay, the time lag linear system and nonlinear system can be on line identified directly. This method can track and detect the variation of the parameters and time delay immediately, and thus can obtain the latest information of the system.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
1999年第4期92-97,共6页
Journal of Southeast University：Natural Science Edition
基金
国家自然科学基金
关键词
时变最优化
信息处理
系统辨识
时变目标函数
time varying optimization
information processing
system identification
