摘要
对工业排放源企业限产是重污染天气应急响应的有效手段.现行限产模式并未考虑扩散条件对限产效果的影响及大气环境治理与经济发展的矛盾.面向特定重污染天气预警,开发数据驱动最优限产方案智能决策方法(DID),用极限学习机预测特定扩散条件下限产方案的减排效果,并将决策者的空气质量改善目标及经济效益目标作为约束对限产方案进行优化.真实数据集上的实验结果表明,在达到同样空气质量目标时,DID方案的经济效益和平均生产比例均高于现行预案;在实现相同经济产值时,DID方案的空气质量改善程度也显著高于现行预案.DID方法能够提高重污染天气应对的精准性及科学性,管理者可以根据当前工作重心设置不同的大气治理及经济发展目标,实现生态环境与区域经济的协调发展.研究结论可为各级政府提高重污染天气应急响应水平提供理论与方法支持,具有一定推广及应用价值.
Production reduction of industrial polluters is one of the most effective measures for heavy air pollution emergency response.Current response methods fail to consider the diffusion conditions and the contradiction between atmospheric regulation and economic development.Addressing those challenges,a data-driven intelligent decision(DID)method is developed.Orienting to diffusion conditions,an extreme learning machine method is applied to forecasting the emission reduction effect of alternative plans,which are then optimized under the two constraints of improving air quality and economic output.Experiments on real datasets demonstrate that the DID schemas are superior to current plans in terms of improving air quality and economic benefit.The DID can give more pertinent and scientific heavy pollution emergency response plans.Managers can set corresponding targets in the DID to balance atmosphere environment protection and economic development.As an effective and low-cost schema,the DID is rational and feasible to be put into practice.
作者
陈富赞
焦扬
李敏强
田津
Chen Fuzan;Jiao Yang;Li Minqiang;Tian Jin(College of Management and Economics,Tianjin University,Tianjin 300072,China)
出处
《系统工程学报》
CSCD
北大核心
2024年第1期1-15,共15页
Journal of Systems Engineering
基金
国家重点研发计划资助项目(2018YFC0213600)
国家自然科学基金资助项目(71771169,72231004).
关键词
决策支持
大气污染
减排
极限学习机
优化
decision support
air pollution
emission reduction
extreme learning machine
optimization
