The Simulation Research Of The Subway Pedestrian Evacuation Model Based On Agent

With the continuous development and improvement of modern urban traffic,the subway takes an increasingly important large passenger transport task.It not only has low energy consumption,little pollution,but also has great significance to the economic development of cities and society.Because of the special nature of its architecture and layout,as well as personnel large densities and high passenger traffic,subway is easy to cause malignant public event such as a large number of casualties and property torts.Therefore,evacuation has become the key and difficult point in the subway safety management.Construction of a reasonable mode and the research of rationality and effective is important and value for evacuation simulating in large public venues.The major works are as following:(1)Based on subway station construction and emergency evacuation environment condition,the factors that influenced the evacuation behavior were analyzed and taking human behavioral psychology as the research direction,the possible psychological and corresponding actions in the event of an emergency were described.Based on the above analysis,the influence of different psychology and behavior on the evacuation process was analyzed emphatically.(2)The evacuation model was designed.In the emergency situation of different people’s mind and behavior as the background and considered the individual behavior and cognitive differences in the emergency evacuation,evacuation of the population was divided into three categories and the Agent technology was introduced into the artificial intelligence theory as a means to construct the evacuation model.Under the Multi-Agent system,evacuation simulation model framework was constructed,and mathematical description of the model was given.The model was consists of a number of different modules,using the theory of cellular automata evacuation geometric modeling and meshing and space environment,in which each evacuated person was regarded as an Agent and each Agent controlled itself in the subway by its own attributes and environment information so that each Agent can walk.(3)The basic particle swarm algorithm is improved.In this paper,the basic Particle Swarm Optimization was improved by using the genetic algorithm theory and applied to the construction of the model.By combining the genetic algorithm and particle swarm optimization,a set of approximate optimal evacuation instruction sequences were found in the instruction of the evacuation instruction sequence,through which Agent combines the physiological and decision to take action and choose a safe exit to escape.Improved algorithm to achieve a static model to a dynamic model transformation,which was compared to basic Particle Swarm Optimization and the simulation results were analyzed.(4)Evacuation simulation program was developed.The evacuation simulation program is based on the established model.The calcuation of evacuation time in Hangzhou Wulin Square Subway Station and others special situation with standard calculation results,it proved that the evacuation model relatively accurate simulate the metro evacuation and combined with the simulation results,suggestions for the daily evacuation of the subway was suggested.