Font Size: a A A

Research On Group Path Planning Based On Multi-agent

Posted on:2020-05-14Degree:MasterType:Thesis
Country:ChinaCandidate:J N ZhangFull Text:PDF
GTID:2416330575961968Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
Group evacuation simulation modeling and path planning are widely used in the film industry,game industry and the train of real life emergency evacuation simulation.This paper is mainly aimed at the evacuation simulation modeling and path planning research of groups in the case of sudden occurrences(such as fires,explosions,etc.)in public places.Simulation of group evacuation has been studied for decades.However,due to the factors such as the psychological state of the person,the diversity of individuals,and the relationship among the groups,the researchers face many difficulties in modeling the evacuation of the crowd and path planning.In many challenges,individual differences,psychological factors and the relationships between people has important influence on evacuation time during the evacuation process,For example,the relatively popular multi-agent crowd evacuation model based on particle swarm optimization(PSO)and ant colony in the past two years fails to reflect the influences of these factors on evacuation time and path.This paper analyzes the psychological and behavioral characteristics of individuals.A multi-agent simulation framework based on emotional state is proposed.A leader-follower model based on emotional state is established on the basis of emotional state modeling,in which the selection of leaders is influenced by emotional state values and experience values.In the path algorithm,because the traditional CSO algorithm has problems such as low convergence precision and local optimality,in order to solve these problems,the sinusoidal inertia weight is used to improve,and the position information of the optimal male individual is added to the male individual,and the young individual is added.In the formula,the factor of the non-mother female individual is added,and the chicken swarm optimization algorithm based on sinusoidal variation inertia weight(WCSO)is proposed.In order to solve the influence of the kinship relationship on the escape path,the improved WCSO algorithm is introduced into the path planning of the simulation model.Then,use MATLAB software to verify the feasibility and effectiveness of the improved algorithm.Finally,the WCSO algorithm is used in the multi-agent model and combined with the corresponding objective function,obstacle avoidance and global optimal update mechanism for path planning research.At the end of the paper,in order to verify the feasibility of the multi-agent model based on emotional state-based leader-follower and the path planning method based on WCSO algorithm.First,Netlogo software was used to simulate single exit and double exit,measure evacuation time and evacuation path of the group,and analyze the reasons.Then,evacuation time of different exit widens in the case of single exit was measured.Finally,it was compared with the typical cellular automata model(CA),social force and multi-agent PSO model.It is concluded that the emotional state-based leader-follower evacuation simulation model based on group relationship and the WCSO algorithm-based evacuation simulation path planning is more realistic in the situation of real evacuation situation.In summary,the multi-agent model and path planning method proposed in the paper have effectiveness and practical significance.
Keywords/Search Tags:Group evacuation, multi-agent, WCSO algorithm, leader-follower, path planning
PDF Full Text Request
Related items