Dynamic Modeling And Complicated Characteristics Of Crowd Evacuation With Consideration Of Information Navigation And Game Behavior

With the development of the urbanization in our country, the material and culture demand of people is growing, public squares, theaters, shopping malls and other public places have become an important part of the people’s daily life. These large public places have the following characteristics:the high dense crowd, the larger fluidity, etc., once the emergency public safety incidents (e.g., fire, trample, etc.) occur, it may induce easily significant casualties and property losses due to improper evacuation behaviors or emergency strategies. Therefore, it is a key link to ensure the large gathered crowds rapid, order and safe evacuation in emergency, and is also a serious problem need be urgently solved. In fact, it has an important meaning to research information navigation, game psychologies and behavior laws of pedestrians in the case of emergency, which will be helpful for scientifically guiding and solving such problems. In this dissertation, based on the lattice gas models and cellular automata models, several improved dynamic models are proposed to investigate crowd evacuation by considering the information navigation, psychological characteristics of pedestrian game and leader effect in smoke situation, respectively. And the dynamic mechanisms of pedestrian evacuation are explored via the numerical simulation and theoretical analysis. Furthermore, we put forward some suggestions for the safety evacuation of pedestrians and the reasonable design of buildings.The main contents of the dissertation are as follows.I. Based on the Tajima’s model and combining the concept of "static field" in cellular automaton model, an extended lattice gas model is established to simulate crowd evacuation in a hall by considering information navigation, and the influences of the numbers, position distributions and strengths of information source are investigated, and the macroscopic phenomena of pedestrian flow induced by the information navigation is also discussed.Based on the original lattice gas model proposed by Tajima et al, and combining the concept of "static field"" in cellular automaton model and introducing "recognition"’, which reflect decision-making behavior after pedestrian learn information, an improved lattice gas model is proposed to simulate crowd evacuation in a hall by considering information navigation. The numerical simulations are performed to investigate the effects of the numbers position distribution and strength of information source on the evacuation process in a hall. And the typical spatio-temporal characteristic during the evacuation process is also discussed. The simulate results show that the evacuation efficiency could be improved obviously when information navigation is considered. The information source at the exit plays an important role in evacuation. In addition, during the emergency evacuation, if pedestrian could move strictly according to the correct information navigation, the whole evacuation time will be decreased drastically and the corresponding evacuation efficiency will be improved greatly.II. A dynamic changes of pedestrian behavioral strategies lattice gas model is proposed to simulate crowd emergency evacuation by considering the game behaviors of pedestrian, and the influence of the proportion of the initial cooperators, the parameters of game proceeds matrix and the strength of drift on the crowd evacuation dynamics are investigated, and the evolution laws of pedestrian movement and formation mechanisms of space-time patterns are also studied.Based on Tajima’s lattice gas model, an extended lattice gas model is established to simulate crowd evacuation by considering the game behaviors of pedestrians, which to solve the problem of position conflict among pedestrians. The influences of the proportion of the initial cooperators, the parameters of game proceeds matrix and the strength of drift on pedestrian evacuation dynamics are studied via numerical simulation. The simulation results show when the game behaviors of pedestrians are considered, it is cannot improve the efficiency of the evacuation, on the contrary, lead to the evacuation process becomes slow, and the evacuation time delay. When the system does not exist in the defectors, evacuation efficiency is the highest. The larger strength of drift, the higher evacuation efficiency. In addition, when game among pedestrians occur during the evacuation process, two people games are dominant, and the larger parameters p of the game proceeds matrix, the higher the evacuation efficiency, and the parameter q has almost no effect on the evacuation efficiency.III. Aiming at smoggy environment caused by fire, an extended Moore neighborhood lattice gas model are established to simulate crowd evacuation by considering leader’s guide, and the leader effect, the underlying mechanism of macroscopic phenomena and dynamic characteristics of space-time are studied.Aiming at smoggy environment caused by fire, the visual field and information of pedestrian are limited, an improved Moore neighborhood lattice gas model are proposed to simulate crowd evacuation by considering leader’s guide. The influences of with or without leader, the size of the region of leader effect, the rational and irrational behaviors of leaders on evacuation efficiency are investigated via numerical simulation. The simulation results show that the rational leaders play an important role to improve the efficiency of evacuation. The greater is size, the higher the evacuation efficiency. In addition, there is smoke case, more people obey the leader’s command and follow their behaviors, the higher the evacuation efficiency.The final chapter of this dissertation is devoted to summary of the thesis and an outlook of further study on the crowd evacuation.