基于细菌觅食优化多核支持向量机的作物生长环境控制  

Crop Growth Environment Control Based on Bacterial Foraging Optimization Multi-kernel Support Vector Machine

作　　者：蔡桂全 陶建平[1] CAI Guiquan;TAO Jianping(College of Economics and Management,Huazhong Agricultural University,Wuhan 430070,Hubei,China;Business School,Yulin Normal University,Yulin 537000,Guangxi,China)

机构地区：[1]华中农业大学经济管理学院,湖北武汉430070 [2]玉林师范学院商学院,广西玉林537000

出　　处：《济南大学学报（自然科学版）》2023年第3期303-308,共6页Journal of University of Jinan(Science and Technology)

基　　金：国家自然科学基金项目(71773033);2020年度广西高校中青年教师科研基础能力提升项目(2020KY14025);2019年度广西民族地区文化建设与社会治理研究中心研究项目(2019YJJD0008)。

摘　　要：为了解决大规模生长环境变量所带来的计算复杂度较高的问题,采用细菌觅食优化多核支持向量机算法对农作物产量进行预测分析,从而实现作物生长环境的最优控制;首先,采用高斯核函数、多项式核函数和Sigmoid核函数组合方式建立多核支持向量机,其输入为作物生长环境,采用细菌觅食优化算法优化核函数关键参数;其次,利用多核函数的参数构建菌群进行训练,设置作物产量作为细菌觅食优化算法适应度;最后,通过菌群位置更新优化后的最优核函数参数进行多核支持向量机优化求解,获得空气温度、湿度,土壤温度、湿度等生长环境特征数据。结果表明,选择合适的多核函数组合,并合理设置细菌觅食优化算法的引力和斥力系数、迁徙概率阈值等参数,能够获得最高产量所对应的作物生长环境特征数据。To solve the problem of high computational complexity caused by large-scale growth environment variables,multi-kernel support vector machine was used to predict and analyze crop yield,so as to realize optimal control of crop growth environment.Firstly,a multi-kernel support vector machine was established by combining Gaussian kernel function,polynomial kernel func tion,and Sigmoid kernel function,whose input was crop growth environme nt,and bacterial foraging optimization algorithm was used to optimize key parameters of kernel function.Secondly,the parameters of multi-kernel function were used to construct flora for training,and crop yield was set as the fitness of bacterial foraging optimization algorithm.Finally,the optimal kernel function parameters after updating and optimizing the flora position were solved by using multi-kernel support vector machine,and the growth environment characteristic data such as air temperature and humidity,soil temperature and humidity were obtained.The results show that the characteristic data of crop growth environment corresponding to the highest yield can be obtained by selecting the appropriate combination of multi-kernel functions and reasonably setting the parameters such as gravity and repulsion coefficients and migration probability threshold of bacterial foraging optimization algorithm.

关 键 词：智慧农业 多核支持向量机 细菌觅食优化算法 生长环境 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]