Crowd-driven Mid-scale Layout Design Based On Real Environment Information

The mid-scale layout such as malls,schools and subway stations,has unique functions,which play an irreplaceable role in urban development.It is of great practical significance to realize the reasonable layout.Humans are the users of these regional layouts and designing from the needs of people is the best embodiment of the people-oriented concept.The layout full of humanity is the basic component of building a livable environment and building a harmonious and beautiful society.In order to highlight the importance of crowd factor in medium-sized layout design and realize the reasonable layout planning,this paper takes the attribute state of crowd in simulation as the evaluation standard of layout,and completes the layout optimization based on generation value driving,and proposes a new idea of crowd simulation driving to complete the task of space layout optimization.The thesis uses GIS map to establish the simulation environment corresponding to the real environment,and constructs a crowd simulation platform integrating the real environment information.The experiment put agent as attribute carrier into the simulation environment,and drive agent movement by social force model.GIS map file can not only provide real and accurate simulation environment for simulation,but also ensure the sharing of entity information in simulation system,and realize the shortest process time and optimal path of the whole crowd simulation.Based on the simulation of the virus spread in the university campus under the epidemic situation caused by New Coronavirus,the virus propagation rule and trend were studied,and the validity of the simulation environment and the platform were verified.Based on the above platform,aiming at the mid-scale layout optimization problem,this thesis proposes a method to optimize the layout of local areas through crowd simulation,and then realize the overall layout planning.This thesis abandons the traditional layout design method which only considers the quantity constraints or uses heuristic algorithm to search the layout that meets the conditions,and pays attention to the close relationship between visitors' feelings when visiting the layout and the planning scheme of spatial layout.Taking the layout of a large shopping mall as an example,this paper proposes to use the layout features to express the layout,and use various costs to express the local advantages and disadvantages of the layout.Through machine learning to train the learning model,the mapping relationship between the layout features and the layout costs is established.By predicting the generation value of each area in the layout,the local area is dynamically optimized to complete the optimization of the overall layout.Finally,the feasibility of this method is verified by crowd simulation.