摘要
应用贝叶斯网络解决地下水环境管理中具有不确定性的多目标决策问题,通过对决策变量氮肥施用量以及灌溉模式的调控,减少水中的硝酸盐含量,达到既能有效改善水环境又不至使农民经济利益受到损害的目标。通过分析具体的地下水环境管理系统中变量间的相互关系,构建描述变量间不确定性关系的贝叶斯网络模型,其中包括表示其依赖关系的有向无环图和表示其具体概率依赖程度的条件概率表。并在多个水环境管理目标均达到最优的前提下进行概率推理,得到决策变量氮肥施用量以及灌溉模式取不同值时目标变量的概率分布情况。最终确定出能使所有目标均达到最优的合理的水环境管理决策：（1）使用喷灌,将氮肥施用量控制在0.01~0.03 kg/m2;（2）使用漫灌,将氮肥施用量控制在0.01~0.02 kg/m2。
Bayesian network was applied in groundwater environment management to deal with the uncertainty of multi-object decision-making problem in this paper.Through regulation of decision variables,i.e.the nitrogenous fertilizer application amount and the irrigation mode,the content of nitrites in water could be reduced,and the objective of improving water environment and not harming the economic interests of farmers would be achieved.The relationships between variables were analyzed,and then the Bayesian network model was constructed,including directed acyclic graphs which describe the dependent relationships of variables and conditional probability tables which express the specific level of the dependency.On the premise that all of the multiple water environment management objectives achieved the intended goals,the probabilistic inference of Bayesian network was drawn,and probability distribution of objective variables at different decision variables value was obtained.At last,reasonable decision support which can make all the objectives achieve the optimal was determined：（1） using sprinkler irrigation,nitrogenous fertilizer application amount should be between 0.01 kg/m2 and 0.03 kg/m2;（2） using flooding irrigation,nitrogenous fertilizer application amount should be between 0.01 kg/m2 and 0.02 kg/m2.
出处
《地学前缘》
EI
CAS
CSCD
北大核心
2010年第6期247-254,共8页
Earth Science Frontiers
基金
国家高技术研究发展计划"863"重大项目(2008AA06A410)
吉林省科技厅科技发展计划项目(20080456)
关键词
贝叶斯网络
地下水环境管理
不确定性
Bayesian network
groundwater environment management
uncertainty
