摘要
非线性回归模型的建立是一类重要的问题,给出了一种基于三角进化算法的求解方法。利用最小二乘法将问题中参数的求解转化为无约束函数优化问题,而后利用种群并行搜索策略的三角进化算法对其求解。此方法不受问题连续、光滑的限制,避免了大量求导的计算,数值实验的结果证明了该算法的全局收敛性和有效性。
Nonlinear regression model generating is a class of important and difficuh problem. Based on triangle evolution, a method for solving regression coefficient is proposed in this paper. The problem is transformed into optimization problem without constraints to be solved by triangle evolution which is a parallel searching strategy. This algorithm, free from limitation of the form of equations, avoids such complicated process as calculating derivative and finding inverse of a matrix and gives the approximate solution or accurate solution. Numerical computations show that this algorithm is convergent globally and effective.
出处
《承德石油高等专科学校学报》
CAS
2009年第1期52-54,共3页
Journal of Chengde Petroleum College
关键词
非线性
回归
进化算法
nonlinear
regression
evolutionary algorithm
