三次设计结合模矢法拟合Logistic曲线的研究

Ｌｏｇｉｓｔｉｃ曲线是种群动态的一个经典模型，针对传统拟合方法存在的缺陷和三次设计的优点，以及拟合过程是离差平方和最小──可看作是运筹学中无约束极值问题──的优化原则，本文提出新的方法－三次设计结合模大法的组合方法来拟合，结果表明该法能取长补短，拟合效果最佳，因此在一般非线性模型参数求解中具有普遍意义．

A Study on Optimum Fitting of Logistic Curve by Combinatorial Method of Three Design and Pattern Search

Logistic curve is a classic model of population dynamics. In view of the shortcomings of the traditional methods to fit it, based on the advantage of three design method and the optimum principle, and the minimizing residuals for the equation regarded as the question of seeking extreme value without constraint conditions in Operational Research, the paper advances a new method, the combinatorial method of three design and pattern search, to fit Logistic curve, whose result shows that it can make up each other's deficiencies and fit to the data best, and so it has a wide application in determining the values of parameters of nonlinear curve.