| With the development of society and the prosperity of economy,the degree of socialization of people’s lives has continued to increase,and gatherings that meet a large number of people have become more frequent.In these public rallies,crowds and tramples can easily occur due to the gathering of thousands of people.Therefore,studying the layout of important elements(such as exits,obstacles,etc.)in the building environment,so that when an emergency occurs,a large number of people can be evacuated from each emergency exit in the shortest time,is an important content that needs to be carefully considered in the architectural design.In order to obtain the obstacle layout with the best crowd evacuation effect,it is necessary to design and implement an optimization method for the layout of important elements in the building environment,to promote the flow of people during the emergency evacuation of buildings,and to ensure the safety of large-scale crowds in public rallies.This thesis conducted an in-depth investigation and analysis of the research status of obstacle layout in the building environment.In order to obtain the obstacle layout with the best crowd evacuation effect,this thesis has carried out research from two aspects: the construction of crowd simulation system for obstacle optimization and the design of obstacle layout optimization algorithm.The construction of crowd simulation system for obstacle optimization is mainly to model the environment and pedestrian movement behavior in the evacuation scenario to simulate the crowd evacuation process,so as to evaluate the safety evacuation performance of the obstacle layout and feed back to the layout optimization algorithm to generate a layout with better evacuation performance.The design of the obstacle layout optimization algorithm is mainly to design an optimization algorithm that can comprehensively consider the impact of the number and attributes of obstacles in the building environment on crowd evacuation to find the most favorable obstacle layout for crowd evacuation,and also design a search space update algorithm to speed up the layout optimization algorithm convergence.The following points are the main contributions of this thesis:(1)A grid index-based environment representation method and a pedestrian motion behavior simulation framework composed of three levels of perception,decision-making,and motion are proposed to build a crowd simulation system for obstacle optimization,which is used to evaluate the safety evacuation performance of obstacle layout and feed back to the layout optimization algorithm to guide it to find a layout with better evacuation performance.(2)An optimization method of obstacle layout based on MA-ES is proposed,in which variablelength chromosome is used to abstract obstacle layout.The optimization of variable-length chromosome by MA-ES optimizes the number and attributes of obstacles in the environment,and finally obtains the obstacle layout with the best crowd evacuation effect.(3)A search space update algorithm is proposed to dynamically adjust the search space during layout optimization,so as to speed up the convergence of the algorithm.(4)A crowd simulation system for obstacle optimization was implemented and tested.Through comparison experiments with other layout optimization algorithms in different scenarios,the effectiveness and superiority of the layout optimization algorithm proposed by this thesis was verified.In addition,the influence of the parameters of the layout optimization algorithm on its performance and the sensitivity of the layout optimization algorithm to initialization were also explored.Finally,the obstacle layout experience in different scenarios was provided. |
