基于IEA优化的农药降解GM(1,1)预测模型

The GM (1,1) Prediction Model of Pesticide Degradation Optimized by IEA

将免疫进化算法(IEA)和灰色系统理论结合起来,建立了农药降解的IEA-GM(1,1)预测模型,分别对喹噁硫磷在豇豆中的残留量、代森锰锌在西红柿中的消解动态、粉锈宁在麦穗中的残留量、抗蚜威在黄瓜果实中的残留量进行预测。结果表明,IEA-GM(1,1)预测模型拟合精度和拟合效果明显优于其他模型,而且该模型不受时间等距条件的限制,建模时不用进行时间变换,可用于预测施药后任意时刻的农药残留量。IEA和灰色系统理论同时用于农药降解建模原理直观、简便、易行,为农药在生态环境中降解规律和降解模型的研究提供了一条新的途径。

The IEA-GM （1,1） prediction model of the pesticide degradation was established according to immune evolutionary algrorithm （IEA） and gray system theory. And the degradation of residual quinalfhosion in cowpea, the degradation of residual mancozeb in lycopersicon esculentum miller, the degradation dynamics of triadimefon in ear of wheat and the dynamic dispelling of pirimicarb residue in cucumber fruit were predicted based on the IEA-GM （1,1） prediction model. The results showed that the IEA-GM （1,1） prediction model exhibited a higher accuracy with a higher correlation coeiticent （R^2） and a lower residual sum of square （S^2） than other models in the corresponding practical ap- plications. According to IEA-GM（1,1） prediction model, the correlation coeiticents （R^2） of the degradation of residual quinalfhosion in cowpea, the degradation of residual mancozeb in lycopersicon esculentum miller, the degradation of triadimefon in ear of wheat and the pirimicarb residue in cucumber fruit were up to 0.999 9, 0.971 6, 0.996 3 and 0.996 3, respectively. Moreover, this model was not limited by the time interval, and it was proved to be simple, convenient and feasible in the calculation of pesticide degradation at any time. Therefore, it would provide a new theoretical way for the establishment of the pesticide residual model and the research of pesticide degradation in ecological environment.