摘要
遗传规划(GP)是一种基于达尔文进化理论的数学规划方法。讨论了GP在符号回归中的应用。与传统的数据拟合方法相比,GP不必给出拟合函数的形式,同时,在初始群体足够大而且交叉和变异概率设置合理的情况下,不会陷入局部优化,具有更广泛的适用性。对于不给定函数形式的曲线拟合,GP可以自动得到曲线的函数形式及其参数大小,避免了传统方法的缺陷。通过具体的应用实例,说明了GP在测量数据处理中的应用。
Genetic programming(GP) is a kind of mathematical programming method based on Darwin's theory of evolution. The application of GP in symbolic regression is discussed. No fitting function form for data fitting is needed while running GP and global optimum can be gotten with reasonable cross and mutation probability. The fitting for a arbitrary curve is impossible with conventional method, while with GP it is easy to implement. Examples are given to explain data fitting of measurement by GP.
出处
《传感器与微系统》
CSCD
北大核心
2007年第11期108-110,共3页
Transducer and Microsystem Technologies
关键词
遗传规划
符号回归
数据拟合
genetic programming(GP)
symbolic regression
data fitting