论文部分内容阅读
为了提高突发事故应急效率,针对事故发展的不同阶段对应急资源的需求,在群体决策方法的基础上利用支持向量机方法进行动态应急预案优选。首先利用专家对量化后的应急预案进行评价,通过向量拼接将其转化为等价的可用于分类的训练样本集,再应用标准支持向量机分类器对应急预案进行分类和排序,最后实现对应急预案的动态优选。研究表明,该方法充分利用了支持向量机方法的自学习能力,以应急过程中的动态需求为依据,可以实现动态的应急预案优选。
In order to improve the emergency response efficiency of emergency, aiming at the demand of emergency resources in different stages of accident development, the dynamic emergency response plan selection based on the group decision-making method is based on SVM. Firstly, experts evaluate the contingency plans after quantification, transform them into equivalent training sample sets which can be used for classification by vector splicing, and then use standard SVM classifiers to classify and sort contingency plans. Finally, Dynamic optimization of the plan. The research shows that this method makes full use of the self-learning ability of SVM method, and based on the dynamic requirements in emergency process, the dynamic emergency plan optimization can be realized.