基于曲线拟合改进算法的水质生物毒性研究

Water Quality Biological Toxicity Research Based on Curve Fitting Improvement Algorithm

水质综合生物毒性种类和浓度的预测通常采用曲线拟合模型。文中提出一种基于数值相近(最小二乘法)与形态相似结合的曲线拟合算法,利用两种算法的精确度值作为改进算法模型中的权重,并引入时间影响因子和powell算法。解决数值相近(最小二乘法)原理的拟合速度,时间因素对发光菌活性影响以及利用权重值优势避免powell算法的前n个搜索方向必须线性无关等问题。研究结果表明,改进的算法提高了运算速度,能反映发光菌活性及避免powell算法无最优解。

Comprehensive types and concentrations of biological toxicity of water quality prediction usually used curve fitting model. This paper presented a curve fitting method based on numerical （The least square method） similar combining with similar shape. The accuracy of the two algorithms were used as the weight of improved algorithm model, and the time effect factor and the Powell algorithm were introduced. The problem of numerical principle of （ The least square method） fitting speed, time factors influence on the luminescent bacteria activity and weight value advantage were utilized to avoid the first n search direction must be linearly independent in Powell algorithm were solved. Research results show that the improved algorithm improves the operation speed, reaction luminescent bacteria activity and can avoid the Powell algorithm has no optimal solution.