基于方波信号的半全局极值搜索算法

Semi-Global Extremum Search Algorithm Based on Square Wave Signal

摘要:

针对半全局极值搜索算法,采用正弦信号作为激励信号时,幅值自适应律初始值选取严格、输出存在颤振的问题,提出了采用方波信号作为激励信号。利用平均化方法证明了激励信号选取较小的初始值即可使目标函数收敛到全局极值,同时减弱了输出颤振；采用正弦和方波信号作为激励信号进行仿真对比,说明方波信号可以提高算法的半全局收敛性,有效减弱输出颤振。

Abstract:

In this paper, a semi-global extremum search algorithm based on square wave signal was proposed to improve semi-global convergence and effectively weaken the output chatter, and the averaging method was used to prove the system can be achieved by the semi-global extreme in present of a smaller initial value. Sinusoidal signal and square wave signal were used as the excitation signal in simulation to verify the validity of the results.