Dissertation > Excellent graduate degree dissertation topics show

Methods and Technologies for Developing Emergency Action Plan Based on HTN Planning

Author: TangPan
Tutor: WangHongWei
School: Huazhong University of Science and Technology
Course: Systems Engineering
Keywords: unconventional event emergency decision making emergency action plan ontology knowledge modeling hierarchy task network planning decision model asynchronous incident objective
CLC: X4
Type: PhD thesis
Year: 2011
Downloads: 91
Quote: 0
Read: Download Dissertation

Abstract


Unconventional events are extreme events which our society has never or rarely undergo, and lack the knowledge of the evolutional law and disposal experience. They often break out suddenly, cause or may result serious harmfulness for our society. Therefore, it is needed to take actions for handling them by the society organizations. Additionally, they challenge the ability of governments and our society for taking effective emergency management methods and measures.In the current time, unconventional events have brought severe shock to economy and society of our country, lead to tremendous economic losses, and cause serious social issues. Meanwhile, they have become the huge hidden risks to build a harmonious society and implement sustainable development strategy in our country. Unconventional events have the properties of widespread impacting areas and high intensity, and involve multiple coupling disasters or multiple incidents which happen simultaneously. All of them lead that the normal methods for preventing and controlling incidents loses their effects. Additionally, the emergency response process cannot be deployed and implemented according to the emergency operation plans of a single agency independently. It needs multiple government departments and social units from different areas and with no collaboration experience to provide emergency services and schedule emergency resources for accomplishing the emergency response work together in the limited time.Developing emergency action plans scientifically and effectively is the basic means to coordinate emergency actions of multiple agencies participating in the emergency response process. It is the central decision problem in unconventional events emergency management. The essence is to develop the emergency action plans for achieving the identified incident objectives facing the complex emergency situations. However, unconventional events emergency response process has the common characteristics, such as dynamic, complexity and time presses. All of them challenge the decision making ability of emergency managers for developing emergency action plans. Additionally, the emergency action plan developing problems are typical complex and unstructured decision problems. The problem solving process involves selecting and restructuring heterogeneous emergency domain knowledge of multiple agencies with no collaboration experience. Traditional mathematics modeling and optimization methods are hard to handle them.Hierarchy task net planning paradigm searches action plans by decomposing tasks. That is similar to the thought process of emergency managers while developing emergency action plans. Therefore, it is an effective method for modeling and solving unstructured decision problems. In this thesis, the decision making process which the emergency command team develops emergency action plans during unconventional events is taken as the research object. Then, focusing on modeling and solving the decision problems closely, the methods and technologies for developing emergency action plans based on HTN planning is studied according to the knowledge modeling theories and automated planning technologies. The main research work includes the following several aspects.The decision maker, decision making objects and decision making environments, and their characteristics for developing unconventional events emergency action plan is analyzed. Then, the research object is defined according to these three aspects. On the basis, a decision making process model is proposed according to relevant literatures of emergency decision making theory. It describes the decision making process and behaviors of emergency command teams for developing emergency action plans during emergency response. The formal definitions of the relevant concepts involving in the decision making process are given. All of them lay the foundation of the following research work.Focusing on the modeling of heterogeneous domain knowledge involving in the emergency action plan developing process, the emergency plan papers is taken as the modeling objects to study the methods for modeling emergency domain knowledge. At present, they are unstructured documents dispersed in multiple organizations, and lack of a unified modeling method. Therefore, it is difficult to represent and utilization the domain knowledge described by them. According to analysis of the main elements and their structural characteristics for emergency plan papers, an emergency plan ontology model is proposed based on reusing the present ontology models. Additionally, an application study is carried out for modeling the flood controlling emergency plan documents of some area.Centering on solving the unconventional events emergency action plan developing decision problems, a set of planning requirements for designing HTN planning paradigms applied in this specific domain is proposed by analyzing the characteristics of emergency response process. Additionally, the shortcomings of current HTN planners are discussed according to the reference framework. Based on the guidance, an anytime temporal HTN planner is designed for assisting emergency command teams to develop emergency action plans. It follows the anytime principle to search action plans, can be interrupted at any time and output the generated plans. Meanwhile, the quality of the plans is improved as more planning time is given. Additionally, the planner is a state based forward HTN paradigm, and implements the search process without backtracking. Therefore, it can imitate the decision logic of emergency managers during developing emergency action plans. Finally, a set of heuristics for selecting search nodes during the planning process is proposed for improving the efficiency of the planning process.An algorithm for translating the ontology knowledge models of emergency plan to knowledge formalisms of our HTN planner is designed. It handles difficulties of knowledge modeling for applying the HTN planner to solve real-world problems. Meanwhile, a flood evacuation domain of some area is designed and a set of experiments is carried out for testing effectiveness and practicability of this planner. Experiment results show that it satisfies all the proposed planning requirements.Finally, a novel emergency action plan decision model for assisting emergency managers to develop action plans is present based on the HTN planner. The decision model follows the integrated planning and execution principle. It executes the plan refinement steps for the tasks achieving the incident objectives with higher priorities during the planning process, and start to execute the plan before it is generated completely. Multiple incident objectives are identified dynamically during emergency response. According to the characteristics, an algorithm is proposed which can handle asynchronous incident objectives with priorities efficiently. Incident objectives with higher priority take precedence over others to be handled and the ones with lower priority are handled opportunistically.

Related Dissertations

  1. The Effect of Instruction for Middle School with Philosophy,G633.6
  2. Murine Peritoneal Macrophages Transcriptional Responses Following in Vivo Infection with Streptococcus Suis Type 2,S858.91
  3. Semantic Retrieval Research Based on Ontology,TP391.3
  4. Lukacs ' ontology of social existence \,B515
  5. Ontology -based Distributed Description Logic Modular Construction Methods,TP391.1
  6. Ontology -based Semantic Web service matching and composition method,TP393.09
  7. WordNet and the \,G254
  8. Latour 's actor-network theory,N02
  9. Research of the Model of Enterprise Competitive Intelligence Collection System Based on Cross-Language Information Retrieval,TP391.3
  10. Research of Text Clustering on Food Complaint Documents Based on Ontology,TP391.1
  11. Research on Mapping Relational Database to RDF (S),TP311.13
  12. Research on Semantic-Based Web Service Discovery,TP393.09
  13. Research on the Evaluation and Model of Apparel Supply Chain Risk for Apparel Manufacturing Enterprises,F274;F224
  14. Domain Ontology-based Marine Environmental Data Warehouse Design,TP311.13
  15. Study on Discovering the Relationships among Data Resources in DataSpace,TP311.13
  16. Philosophical Research on the Intelligence Space in the Pervasive Computing Environment,N02
  17. The Research for Strategy of Enterprise-level Software Components Quality Assurance and Rapid Deployment,TP311.53
  18. Study on Semantic Model of Digital Emergency Plan Under the Scenarios-Cope Mode,TP391.1
  19. Ontology-based medicine named entity recognition technology research,TP391.1
  20. Ontology-based prescription recommend system model and application,TP391.3
  21. Ontology-based Knowledge Retrieval System Smart Grid,TM76

CLC: > Environmental science, safety science > Disasters and their prevention
© 2012 www.DissertationTopic.Net  Mobile