基于GA-BP神经网络的雨生红球藻生长趋势预测

Prediction of Growth Trend of Haematococcus pluvialis Based on GA-BP Neural Network

虾青素具有极强的抗氧化功能,然而在自然界中,能够在相同时间内积累最多虾青素的生物主要为雨生红球藻。但其生长周期较长且对生长环境依赖性强,因此其生长环境参数与其生长、繁殖状况之间存在着非线性关系。BP神经网络模型运用了梯度下降法的基本思想,可以较好逼近复杂的非线性关系。但由于BP神经网络容易陷入局部最小,从而影响预测的结果,所以采用遗传算法对BP神经网络进行优化。根据影响雨生红球藻生长的主要环境(pH),建立基于遗传算法的BP神经网络模型,并对雨生红球藻生长状况进行试验验证,准确率可以达到90%以上。

Astaxanthin has a strong antioxidant function.However,in nature,the organisms that can accumulate the most astaxanthin at the same time are mainly Rhodococcus pluvialis.However,its growth cycle is long and strongly dependent on the growth environment,so there is a nonlinear relationship between its growth environment parameters and its growth and reproduction.BP neural network model used the basic idea of gradient descent method,which could better approach the complex nonlinear relationship.However,because BP neural network was easy to fall into local minimum,which affected the prediction results,genetic algorithm was used to optimize BP neural network.According to the main environment(pH)affecting the growth of Haematococcus pluvialis,a BP neural network model based on genetic algorithm was established,and the growth status of Haematococcus pluvialis was experimentally verified.The accuracy could reach more than 90%.