Research On The Model Of Emergency Decision Making For Dangerous Goods Based On Bayesian Network

Emergency decision is one of the key problems in the research of emergency management in recent years. It is one of the leading issues in the field of emergency management. Dangerous goods emergency rescue due to accident harm violently, causing heavy losses, complexity of the rescue environment and respond to the special requirements of professional processing, coupled with concerns casualties rescue, cordoned off the area, endanger mass evacuation, field control decontamination, and environmental monitoring data, public opinion guidance information such as a series of action, is a very complex decision problems and practical problems. Therefore, it is very important to carry out research on the emergency decision of the dangerous goods. Traditional emergency decision making pattern dependence to accumulate experience and subjective judgment has been in dangerous goods emergency rescue decision process demonstrate the limitations, could not adapt to the complex dynamics of dangerous goods emergency, need to expand and to decision model of innovation. The purpose of this paper is to make a quantitative analysis of the emergency decision of the dangerous goods, through the establishment of a Bayesian emergency decision model, identify the key factors that affect the emergency decision, and provide support for emergency decision making.Decision making is a complex decision problem, and its influential factors are numerous, the identification of effective factors and its impact mechanism is an important prerequisite for the study of emergency decision making. Through access to a large number of dangerous goods related rules and regulations and accident investigation report, from planning, personnel, environment and technology in four aspects: the extraction and identification of the 28 factors, respectively each aspect of dangerous goods emergency decision has carried on the thorough analysis. On this basis, based on the Bayesian network theory, the final 23 factors are identified as the network nodes. Through access to 33 cases of dangerous goods accident investigation report and introduced the case, access to the machine learning sample database, with K2 algorithm Ge NIE2.0 software, in the editing background knowledge of network structure based on the import sample database for the structure learning, combined with expert knowledge to optimize the network, according to statistics of the accident investigation report to obtain the conditional probability of the root node, using the software parameter function and Realization of the parameters of the neural network learning, thus completing the dangerous goods emergency decision model is constructed.Based on an example, the paper introduces the application of network reasoning evidence information analysis method, and through the 33 cases of dangerous goods accident investigation report evidence information update reasoning to construct the network to carry out the rationality verification. The sensitivity analysis function of Ge NIE2.0 software was used to find out the 12 important factors that affect the emergency response of the dangerous goods, and combined with the actual identification of the key factors of the emergency decision chain. The construction of the whole network, there is a solid theoretical basis and machine learning data, to provide the necessary decision support for the emergency decision of dangerous goods.