基于前向PD控制的拟态物理学优化算法研究

Adaptive Artificial Physics Optimization with PD Controller

针对拟态物理学算法局部搜索能力弱的缺陷,提出一种在拟态物理学算法模型中前向引入PD控制器的改进算法,通过对APO算法的位置迭代公式化简后进行z变换,并前向引入PD控制器,再通过z反变换,并化简得到位置的迭代公式和速度的迭代公式,从而得出APO的改进模型APO-PD,最后通过低维和高维的函数性能测试说明了改进的APO算法的有效性。

To solve the disadvantage of poor local searching ability, an improved on the forward PD control through introducing the PD controller as a forward to optimization was presented based the APO （APO-PD）. By carrying out z-transformation on the simplified algorithmic model, and introducing the PD controller into the front path of the model, then carrying out z-l-transformation on the new model, and simplifying the algorithmic model, a APO-PD was got. Finally,the relative function tests in low dimension and multi-dimension show that the APO-PD performs effectively.