Research On Personnel Modeling And Simulation For Public Security Events

In recent years,the international security situation is becoming increasingly complex,and the frequency and scale of public security incidents are rising.During the process of security event,emergency plans and personnel deployment schemes are too subjective,lack of quantitative research and lack of pre verification,which make them difficult to achieve expected effect.It is helpful to improve the effectiveness of event evolution simulation with the research of behavior pattern of different types of people.It can also improve the scientificalness of the plan and enhance the patrolling effect when introducing mathematical model into the design of the deployment and patrol scheme.In this thesis,autonomous agent is used to build human object behavior model under the background of public security event.Personal decision is based on behavior tree model.Different behavior decision logics are constructed for different types of people,such as security guards and inspects.Event triggering model is adopted in personal perception.Different perception modules such as visual and auditory sense are designed based on their different characteristics.Different steering behaviors such as seek,pursue and evade are designed to achieve a natural way of personal movement.The calculation planning module is based on the A*algorithm.The execution efficiency and path planning effect are improved to realize the path planning function of the personnel object.Finally,the Unity3D engine is used to build simulation platform to realize autonomous agent behavior model.The autonomous agent model is used to model personnel behavior,and the interactive relationship different people is rehearsed by the way of computer simulation.The randomness and complexity of the simulation are increased,and the authenticity of the assessment is improved.Unity 3D engine dramatically improves the visualization of dynamic rehearsal display,and enables the user to observe and make decision of public events rehearsal in a realistic three-dimensional simulation environment.In this thesis,the maximum coverage model and the set cover model in combinatorial optimization are used to describe two types of problems in the static deployment of security guards into mathematical problems.Genetic algorithm is used to solve these two models.The dynamic patrol problem is transformed into static deployment problem by using the discrete time slice segmentation method,and the dynamic patrol scheme is designed based on depth first search algorithm combined with the coverage rate threshold pruning strategy.In the process of experiment,the real geographic information obtained from Baidu map API is used to verify the feasibility and compare the effectiveness of the solutions.The experimental results show that the models and solution method proposed in this thesis can solve the design problem of deployment and patrol scheme more efficiently,and improve the scientificalness of the scheme design.The design of the deployment and patrol scheme and the pre simulation verification can improve the effectiveness of security deployment and emergency plan design.It has a certain reference value for the decision-making and implementation of public security incidents.