基于改进D-S理论与SVM后验概率的温室智能控制决策  

RESEARCH ON INTELLIGENT CONTROL DECISION METHOD OF GREENHOUSE BASED ON IMPROVED D-S THEORY AND SVM POSTERIOR PROBABILITY

作　　者：卜娜娜 刘新良 李国民[1] 朱代先[1] Bu Na’na;Liu Xinliang;Li Guomin;Zhu Daixian(School of Communication and Information Engineering,Xi’an University of Science and Technology,Xi’an 710054,Shaanxi,China)

机构地区：[1]西安科技大学通信与信息工程学院,陕西西安710054

出　　处：《计算机应用与软件》2022年第7期20-25,共6页Computer Applications and Software

基　　金：国家自然科学基金青年科学基金项目(61801371);陕西省教育厅专项科研计划项目(15JK1463)。

摘　　要：温室环境中多参数间相互耦合,针对传统温室调控方法的决策因子单一而导致决策片面及精度低的问题,提出利用温室环境中多变量参数特征信息,将权值修正D-S理论与后验概率SVM结合的决策融合方法。利用后验概率SVM构造D-S理论的关键参数基本概率赋值函数(BPA),结合参数间的支持度、相关度和可靠度进行D-S权值修正,利用修正后的权值对BPA进行加权融合处理,根据融合结果对温室环境进行及时调控。实验结果表明,该方法可以准确调控温室环境,有效降低决策的不确定性,显著提高决策的可信度及收敛性。In the greenhouse environment, the multi parameters are coupled with each other. In view of the single decision-making factor of the traditional greenhouse control method, which leads to one-sided decision-making and low accuracy, a decision fusion method is proposed, which combines the D-S theory of weight correction with the posterior probability SVM by using the multi variable parameter characteristic information in the greenhouse environment. The basic probability assignment function(BPA) of the key parameters of D-S theory was constructed by posterior probability SVM. The D-S weights were modified by combining the support, correlation and reliability of the parameters. The BPA was weighted and fused by the modified weights, and the greenhouse environment was timely regulated according to the fusion results. The experimental results show that this method can not only control the greenhouse environment scientifically, reduce the uncertainty of decision effectively, but also improve the reliability and convergence of decision significantly.

关 键 词：温室环境调控 D-S证据理论 权值修正 SVM后验概率 多源数据融合 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]