A Study Of Emergency Decision Knowledge Model Based On CommonKADS Method

Along with society's development in an all-round way and globalization, the modern society has entered into the phase of frequent crisis erupted repeatedly in recent years, the emergency does great harm to people's life and property and social security. How to avoid as far as possible to emergencies happen and minimize the number of emergency brings loss has become the major problem to be solved.In the process of dealing with emergency, emergency decision is the important link of emergency management. Quickly put effective emergency decision-making can reducing the loss of disasters and casualties. To cope various types of incident is a very complicated task, involving high efficiency processing of huge amounts of information, emergency decision-making quick, and so on. However, in the situation that mass information being real-time updated, incident's complexity and uncertainty make the accuracy of the knowledge and efficiency of decision generation hard to be guaranteed. In emergency decision-making process, the environment and the necessary knowledge are constantly change, how to sharing and reuse the existing decision background knowledge to solve new problems is a important way to improve emergency decision-making ability.Emergency decision-making is a knowledge intensive task, According to this characteristic, based on the CommonKADS method, this paper proposed in emergency decision domain knowledge model construction. CommonKADS is a systematic method of knowledge engineering field. It involves knowledge acquisition, knowledge representation, knowledge apply and processing. The characteristic of this method is that it put forward in the "knowledge model" concept. In the knowledge model, knowledge is structured to be divided into three layers, and is easy to realize the reuse.According to the method, we layering three levels to describe emergency decision knowledge:domain knowledge, task knowledge and reasoning knowledge. The domain knowledge used to describe some static knowledge structure of this field, including the domain mode (concept, relationship and rules type) and knowledge base, it contains all aspects of the knowledge that emergency decision-making needed, and it is basis of the knowledge model application. Task knowledge is similar to people's daily life saying "mission", the difference is that task knowledge contains methods to finish the task, In other words, task knowledge described what is the goal of the model, and using what kind of strategies to achieve the goal. Tasks can be hierarchically decomposed into the subtask or reasoning knowledge. Reasoning knowledge is the basic unit of the information processing, and the lowest level function decomposition, it describes the basic steps to reason by using the application domain knowledge.However, the knowledge model that this method eventually constructed is a set of charts and forms, not only hard to operate and maintenance, but also hard to understand and use by computers. Therefore, on the basis of the emergency decision knowledge model that constructed by CommonKADS method, this paper using the ontology method to represent emergency decision-making domain knowledge. Ontology is a "formal and clearly specification of shared conceptual model", it describes the intrinsic property of entities in the field, including the structure of the concept and the relationship between the concepts, and so on. Using ontology method represented domain knowledge could not only greatly increase the maintainability and maneuverability of model, but also effectively conjunct the three levels of knowledge model. It lays a foundation for knowledge model further developed into emergency decision support system.After complete the construction of knowledge model, the usability of the model need to be focused on, that is, whether knowledge model can be used in emergency decision areas to solve knowledge reuse problem. So, in the last chapter, we verified the availability of knowledge model by using some method. Validation methods include verifying the reasoning knowledge by paper simulation validation and use the semantic web development tools-Jena and SPARQL inquires language to reasoning. Thus confirmed that emergency decision knowledge model has certain theoretical significance and application values.