| With the rapid development of China's economy and society,various emergencies have frequently occurred with seriously threatening the safety of people's lives and property.Aiming to the severe condition,a scientific and effective decision-making system for emergency action plans must be established.In the event of a disaster,the system can assist emergency decision-makers to quickly formulate reasonable incident action plans,so as to intervene the development of emergencies as soon as possible to mitigate the damage caused by emergencies.Therefore,the paper focuses on the characteristics of high coupling between actions and resources in emergency rescue.Based on intelligent planning and scheduling methods,a distributed decision-making algorithm for emergency action plans considering resource linkage is established to accelerate the formulation of emergency action plan.After analyzing the existing descriptive and prescriptive emergency action plan decision-making methods,the paper find that these methods can help decision makers to make reasonable emergency action plans to a certain extent,but there are many problems,such as: hidebound patterns,which is difficult to deal with complex and changeable emergencies;complex calculations,which is difficult to deal with the temporal constraint of emergency;emergency decision-makers are under great psychological pressure,which is difficult to deal with the serious consequences of emergencies.In order to overcome the shortcomings of existing decision-making methods,the decision-making algorithm of the incident action plan is designed based on Hierarchical Task Network(HTN)planning method which can deal with the complex domain knowledge and is similar to the reasoning process of the incident action plan.However,the existing HTN planning has a weak ability to reason about resource information and cannot deal with resource linkage in the distributed environment.Moreover,the form of centralized planning cannot be adapted to actual emergencies.The geographically dispersed multiple emergency command teams need to perform task planning at the same time,and the resource linkage problem needs to be attended by multiple agents.To make up for the above shortcomings,the paper combines Distributed Constraint Satisfaction Problems(DCSP)and HTN planning.The basic framework of incident action plan is generated through HTN planning,which includes all actions except the resource linkage plan,and summarizes the unresolved resource linkage problems in the incident action plan.Then the resource linkage problem is uniformly coded into a distributed constraint satisfaction problem,and the improved Asynchronous backtracking(AB)algorithm is used to comprehensively consider the priority of each decision-making agent,the resource shortage state,the resource allocation distance,and the relationship between the upper and lower levels.Furthermore,the complete incident action plan with resource linkage information and resource shortage report is obtained.The method can utilize the powerful reasoning ability of HTN planning,and can exert the advantages of DCSP to solve distributed problems,and improve the AB algorithm to adapt to the actual characteristics of emergency rescue.It is suitable for solving the distributed decision-making problem of incident action plans characterized by information distribution,demand dynamic change with environment,and limited resources.Finally,the above algorithm is implemented using the multi-agent development platform JADE and the hierarchical task network planner JSHOP2.It is verified by various possible scenarios in urban waterlogging.The simulation depicts that the method has advantages in flexibility and processing speed,and has great application value.At the same time,it also enhances the original HTN planning method's ability to deal with distributed problems,enhances its resource reasoning ability,and provides a new idea for the combination of planning and scheduling,which has great theoretical significance. |
PDF Full Text Request