Agent-groups Simulation Technology-based Applications

Massive crowd simulation is a technology which applies Computer Virtual Reality, Artificial Intelligence and Computer Graphics technology to simulate the behavior set of groups which has a large number of individuals. Computer 3-dimensional real time animation of crowds can be applied to training and deployment of military, public safety training and group behavior research. The crowd simulation, which is based on Agent, has overcome the traditional methods'technical limitations. It has already set up the expression of group behavior and individual interaction intelligent model in which the individual psychological, emotional and other non-quantifiable attributes of the crowd (especially the throng) can achieve a great implementation.In this dissertation, we present a new agent-based group behavior model, which propose a set of algorithms or solution in the perception crowd modeling, the behavior crowd modeling and motion decision. In the perception crowd modeling, we propose the division of static environment and dynamic environment perception, focus on and achieve a variety of optimize processing in different circumstances. In the behavior crowd modeling, we propose four sub-components to form a behavior system, and provide some corresponding algorithm for supporting. Based on the perception crowd modeling and the behavior crowd modeling, we propose decision algorithm, which introduces a character reference model parameters"TSVC"about decision-making. The model can excellently transform the intelligent individuals intention of the group。In the motion decision,we can determine the group motion route and behavior patterns from a number of factors such as the goals, status, environmental barriers and dynamic individuals. In addition, we propose the improved dynamic potential field method of calculating the global path and the navigation technology based on test.At the same time, in order to reduce the computational complexity, this paper presents two groups of sub-strategies: the dynamic grouping of the geographical location information based on environmental systems and the dynamic sub-group method based on target points inter-region routing environment.After molding above algorithm model, this paper presents the architecture based on the 3D graphics engine, which is independently developed by our laboratory. Based on the 3D engine, we developed a group behavior simulation module subsystem in the form of plug-in.Finally, we build a 3D crowd simulation environment, which verifies our algorithm models and ideas we built. Then, we achieved satisfactory results.