基于离散Hopfield神经网络的突发事件连锁反应路径推演模型

An emergency event chain reaction path deduction model based on discrete Hopfield neural network

突发事件及其引发的连锁反应会造成严重的灾害后果,如果事先知道突发事件可能引发的连锁反应路径,就可以提前做好断链减灾工作.因此,提出一种基于Hopfield神经网络的突发事件连锁反应路径推演模型.该模型用Hopfield神经网络表示一般的突发事件网络,用Hopfield神经网络的运行规则表示突发事件连锁反应的原理,并设置神经元阈值,将突发事件连锁反应路径的推演过程映射为Hopfield神经网络的演化过程,通过运行Hopfield神经网络推演初始突发事件的一条连锁反应路径.最后,用该模型推演了地震引发的连锁反应路径,验证了该模型的合理性.

Emergency events and the resulting chain reactions would cause serious consequences, if the chain reaction path was known in advance, the chain would be interrupted and the disaster reduced. To solve this problem, an emergency event chain reaction path deduction model based on discrete Hopfield neural network is presented. The model uses the Hopfield neural network to represent a normal emergency event network, then deduct the chain reaction path of the emergency event according to the principle of the chain reaction, and by setting the threshold of the neurons, it maps the inference process of emergency event chain reaction to the evolution process of Hopfield neural network, and the chain reaction path of initial emergency event can be found out by running the Hopfield neural network. Finally, the rationality of this method is validated by analyzing the resulting chain reactions of earthquake.