面向多部门不确定性资源需求的应急任务规划模型

Emergency Task Planning Model for Multi-departmental Uncertain Resource Requirement

大规模突发事件通常需要多个应急响应部门参与,且伴随着大量资源需求,需要为地理上分散的各应急响应部门创建多部门协商应急任务规划模型;同时由于任务规划中各规划主体之间存在信念偏差,所提出的方法考虑了资源不确定性。采用GOCO协议对各规划主体之间的关系进行描述,基于此扩展HTN概念模型以适应多主体参与的分布式规划系统。通过改进传统的OTD算法,设计了能在规划过程中实现多部门协商制定应急任务规划方案的MA-OTD算法。最后,将本研究构建的模型与算法运用到新冠肺炎疫情防控期间湖北省分区分级人员管控问题中,证明该方法的有效性和实用性。为分布式智能规划方法的发展提供了可靠途径,在规划过程中考虑资源不确定增强了HTN规划的应用性能,为现实应急响应工作提供了科学指导。

In view of the fact that large-scale emergencies require lots of resources and the participation of multiple emergency response department,a multi-department negotiation emergency task planning model needs to be created for those departments geographically dispersed.At the same time,due to the belief bias exiting among the planning agents in task planning,resource uncertainty is taken into consideration in the proposed method.Based on the description to the relationship among the planning agents,which uses GOCO protocol,the HTN conceptual model is extended to adapt to the distributed planning system with multi-agent participating.The MA-OTD algorithm that can realize multi-department negotiation.to formulate the emergency task planning solution is designed by improving the traditional OTD algorithm.Finally,the model and algorithm are applied to the management and control of classified personnel in Hubei Province during the prevention and control of the COVID-19,and the effectiveness and practicability of the method is proved.This method provides a reliable way for the development of distributed AI planning methods.The consideration of resource uncertainty during the planning process enhances the application performance of HTN planning,and provides scientific guidance for practical emergency response work.