基于马氏决策过程的突发传染病防控决策模型  

Decision making model of emerging epidemics control based on Markov decision processes

作　　者：于振宇[1] 韩传峰[1] 孟令鹏[1] 

机构地区：[1]同济大学经济与管理学院,上海200092

出　　处：《系统工程学报》2016年第3期338-349,共12页Journal of Systems Engineering

基　　金：国家自然科学基金资助项目(91024023;91224003);中国博士后科学基金资助项目(2014M551459)

摘　　要：研究防控突发传染病在人群接触网络上传播的动态决策问题,在时间维度扩展传染病接触网络模型,真实描述疾病的时空传播规律.综合考虑免疫与隔离措施,基于马尔可夫决策过程构建突发传染病防控动态决策模型,动态生成防控方案,最优化决策周期内的整体决策效果.以典型情景为例建模并仿真求解,验证了模型的可行性和有效性;求解模型得到的动态决策方案可有效抑制疫情规模,为突发传染病防控提供决策支持.Making dynamic decisions according to the spread of emerging epidemics on the contact network is an effective way of epidemics confinement.A contact network model of epidemics is expanded with time dimension,and considers the spread law of epidemics both in time and space in reality.Considering the immunity and quarantine measures,the dynamic decision making model of epidemics is formulated based on Markov decision processes.The model generates control policies dynamically and optimizes total utility over the decision horizon.The feasibility and effectiveness of the model are verified through a typical scenario modeling and simulation,and the results show that the dynamic control policies generated by the model have good performance in containing epidemics spread and can be used as decision support tools for responding emerging epidemics.

关 键 词：突发传染病 接触网络 马尔可夫决策过程 动态决策 免疫 隔离 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]