Multi-source Heterogeneity Information Fusion Methods Based On Ontology Model

How to deal with the disaster is a complicated decision management process after the disaster. The important decision problems-implementation of what kind of rescue operations and how to implement them are in the disaster information collection on the basis of the analysis. Generally, mergency information has the characteristics of multi-source, heterogeneity and uncertainty, so the emergency decision is carried out according to the information of these characteristics. How to more effectively use the information for better decision making result and improve the effect of aid, are the current main problems to be solved in disaster emergency management.Disaster response has the characteristics of multi-stage and multi-agent participation. Emergency rescue process can be divided into multiple stages. The main missions are different in dissimilar stages, so are the needed disaster information types and processing methods. The processing method of disaster information should be based on the specific decision requirements. Based on this idea, this dissertation researchs the emergency response at different stages with different tasks and determines the disaster information fusion methods of multi-source heterogeneous information. This dissertation chooses evidence combination theory and ontology theory as the theoretical basis of research, focus on the disaster situation assessment stage and multi-agent task planning stage. The multi-source information fusion method mainly aims at the two stages of the decision-making target.Firstly, it makes sure the main goal in the situation assessment stage is to evaluate and predict the disaster risk and trend. At this stage, the disaster data has multi-source, heterogeneity and incomplete characteristics. So determining the goal of information fusion is to use the multi-source information complementary characteristics, transform the incomplete, defective, unreliable and conflicting emergency information into the consistent interpretation and description for the target or phenomenon. As the first step, the dissertation puts forward the framework of information fusion system and makes sure the main process of information fusion, so as to integrate the data of different sources. According to the heterogeneous structure of information, this study builds information flow ontology model. Eextracting feature attribute through the knowledge base, Transferable Belief Model(TBM) is secetled as the basic method of data fusion. On the basis, the reliability updating algorithm is put forward on the basis of information timeliness problems. According to multi-objective evaluation problem, the multiple recognition framework fusion algorithm and applied example are put forward.Secondly, decision making is the main task of the emergency action plan during the emergency response mission planning. In this stage, the main problem is the heterogeneous information processing for the distributed and multi-agent joint planning. According to different subjects in different task responsibilities of contingency planning, the agents are divided into three main decision-making levels. The information and processing demand are analyzed for all the levels. On this basis, this study puts forward the main system to cope with the task of joint planning system, and constructs the emergency planning model based on the ontology expression. This model is the basis for emergency response information interoperability between the main agents. Based on this model, the realization algorithm of the emergency plan fusion is put forward as to finish the joint mission planning system of multi-agent. And the application example is analyzed.Thirdly, this study researches the construction of the emergency knowledge base supporting the above functions. From the analysis of emergency information fusion processing the required level of knowledge, this dissertation analyses the structure of knowledge base and content for emergency information base. The emergency information supporting the fusion function is divided into four layers, and this knowledge base is divided into three subdivision bases the ontology base of emergency concepts, ruling base of information fusion and emergency standard process base. The ontology base of emergency concepts stores the basic concepts of emergency fields. This dissertation builds the ontology model of emergency information field as the connotation type knowledge for domain knowledge base. The ruling base of information fusion stores the rules supporting the fusion realization. This study discusses the forms of rule expression. The expression way of emergency action basic units are analyzed in the emergency standard process base.