基于蚁群算法和支持向量机的节水灌溉技术优选  

Optimization of water-saving irrigation technology based on ant colony algorithm and supporting vector machine

作　　者：翟治芬[1,2] 严昌荣[1] 张建华[3] 张燕卿[1] 刘爽[1] 

机构地区：[1]中国农业科学院农业环境与可持续发展研究所,北京100081 [2]农业部规划设计研究院,北京100125 [3]中国农业科学院农业信息研究所,北京100081

出　　处：《吉林大学学报（工学版）》2013年第4期997-1003,共7页Journal of Jilin University:Engineering and Technology Edition

基　　金：'十二五'国家科技支撑计划项目(2012BAD09B01);'973'国家重点基础研究发展计划项目(2012CB955904);世界银行项目(TF 092393-CN)

摘　　要：综合考虑了生产、生态、经济、气象、社会和土壤等因素,建立了节水灌溉技术优选指标体系,利用蚁群算法实现指标的筛选,并以支持向量机为分类器,建立了节水灌溉技术优选模型。以山西省的43个县为案例对该模型进行了试验,试验结果表明,在指标筛选方面,蚁群算法的应用有效减少了指标数量,从初始节水灌溉技术优选指标体系30个指标中,小麦优选出12个指标,玉米优选出16个指标,大豆和棉花优选出17个指标;在节水灌溉技术优选方面,本文模型针对小麦、玉米、大豆和棉花4种作物分别优选出了相应的节水灌溉技术,与当地的实际情况基本吻合。该模型可为决策人提供科学依据,对节水灌溉项目规划设计中选择适宜的节水灌溉技术有较大的现实意义。After considering the factors of production, ecology, economics, meteorology, society and soil, an index system for selecting water-saving irrigation technology was established. An optimization model of water-saving irrigation technology was developed based on Ant Colony Algorithm （ACA） and Support Vector Machine （SVM）. The ACA was used to select indicators and SVM was used to build the classifier. Forty-three counties in Shanxi Province were taken as cases to test the model. Results show that ACA reduces the number of indicators. Form 30 indicators of the index system of water-saving irrigation technology, 12 indicators were selected out for wheat, 16 indicators for corn, 17 indicators for soybean and cotton. The model was used to optimize water- saving irrigation for wheat, corn, soybean and cotton fields respectively, and results were basically consistent with local conditions. The proposed model could provide scientific basis for decision-makers, and has great practical significance in selecting suitable water-saving irrigation technology for planning and designing irrigation projects.

关 键 词：农业工程 节水灌溉技术 优选模型 蚁群算法 支持向量机 

分 类 号：S275[农业科学—农业水土工程]