Research On Key Technologies Of Emergency Decision-making For Unconventional Emergency Based On Scenarioanalysis And CBR

Unconventional emergencies show the unconventionality, with obvious infrequence and inscrutability, highly derivative and concatenate dynamic, typical catastrophic consequences and cannot be controlled, inducing serious social panic and crisis. Unconventional emergencies cannot be predicted in advance based on knowledge and experience in past so that a great deal of difficulties were brought to the government's emergency decision-making and management, and the traditional management pattern of'predict-pesponse'can not meet the demand of unconventional emergency management any more. On the basis of analyzing the system structure and mechanism of emergency decision-making for unconventional emergencies, the key technologies of'scenario-response'emergency decision-making were researched based on scenario analysis method and case-based reasoning in this paper. Main contents were as follows.①On the basis of analyzing the definition and characters of unconventional emergencies, the system dynamic model was established based on entropy and dissipative structure theory, and the entropy changes were analyzed in the process of unconventional emergencies'occurrence, development, evolution, and emergency management. Then the system structure, temporal framework and mechanism of emergency decision-making were analysed on view of mechanism methodology. Dalian oil pipeline explosion accident in 2010 was taken as an example, the content and features of principle, theory, process and operation mechanism were described. The research show unconventional emergency was a dynamic system which was consisted of increasing entropy factors ZS, decreasing entropy factors FS and the system capacity C the early warning and emergency management of unconventional emergencies had to intensify decreasing entropy factors, control increasing entropy factors to form a good dissipative structures, and make the system be located in the orderly balanced and ideal stable area as soon as possible.②According to the established system dynamic model of entropy for unconventional emergencies, the standard swallowtail catastrophe model of unconventional emergencies was obtained. Unconventional emergencies'occurrence mechanism and the phenomena of catastrophe, hysteresis, multi-mode, multi-path in the process of unconventional emergencies'occurrence and management were analyzed by showing swallowtail catastrophe bifurcation set and three-dimensional equilibrium surface chart, and that research show catastrophe theory could explain well the occurrence and control mechanism of unconventional emergencies. Then the catastrophe early warning model was proposed based on triangular fuzzy number theory. The verification of early warning method was implemented on specific example, and the results were consistent with the actual situation of the oil depot. The results of early warning was proved better solidity and the method was shown better reasonable and feasible by analyzing different decision makers'evaluation results (λ, a with different values).③On the basis of defining the'scenario'and analyzing the scenario evolution of unconventional emergencies, the critical works of unconventional emergencies'scenario analysis were definited as representing scenario knowledge, constructing scenario network, and deducting the scenarios. The'Pressure-State-Response'representation structure of scenario knowledge for unconventional emergencies was set up based on PSR model, then the scenario analysis network and deduction model was established based on Bayesian networks theory and the combination rule of D-S evidence theory. The major coal mine gas explosion which happened on mining heading face was taken as an example to verify the scenario analysis and deducting model. The results were consistent with the reality and could provide more information for emergency decision-making and rescue. This method could represent completely the scenario knowledge in the process of scenario evolution, and the qualitative and quantitative scenarios deduction process was achieved.④The case-based reasoning emergency decision-making method of unconventional emergencies based on scenario retrieval was researched, and the improved scenario retrieval algorithm was verified by calculating the similarity between the output of scenario deducting case and specific examples. The research show that the improved K-nearest neighbor approach could improve the effectiveness and accuracy of scenario retrieval by introducing the calculation method of structural similarity as well as the heuristic weights algorithm based on cloud model. By combining scenario analysis method and case-based reasoning, a'scenarios-response'intelligent decision-making platform framework was constructed based on multi-agent system, then the main agents'functional structure and operating mechanism were designed. The study provided a basis to realize'scenarios-response'intelligent decision mode of unconventional emergencies.The research were enriched and expanded on occurrence mechanism and early warning method for unconventional emergencies, and the techniques of emergency scenario deduction were developed. It promoted to achieve'scenario-response'emergency decision-making mode of unconventional emergencies, with great reference value and practical significance to prevent and control unconventional emergencies.