Knowledge Reduction Model of Crowd Evacuation Stability Based on Rough Set Theory

Abstract

To reduce the redundant knowledge discovered from the evolution of crowd evacuation in case of emergencies, this paper aims at knowledge reduction model of crowd evacuation stability evolution. First, the evolutionary mechanism of evacuation stability state is analysed from the perspective of knowledge engineering, and the characteristics of evacuation scenario elements such as physical properties of crowd, structure of evacuation scenario and psychological behaviour of evacuees are analysed and extracted by using self-organizing mapping network algorithm; Secondly, the evolutionary characteristics of evacuation scenario elements and evacuation stability state are discretized and mapped to conditional attributes and decision attributes in rough set universe respectively. Thereby a knowledge representation method of evacuation stability evolution mechanism based on rough set decision information table is constructed. Then, the improved attribute reduction algorithm based on discernibility matrix is used to reduce the knowledge of decision information table processed by self-organizing network algorithm. Finally, a knowledge reduction model of crowd evacuation stability based on rough sets is proposed. It is validated effective to achieve the decision-making of crowd stability, and which provides theoretical basis and methodological support for scientific guidance of emergency evacuation.