| COVID-19 is highly pathogenic and infectious,requiring adaptation of prevention and control strategies and containment of the epidemic’s spread.The use of multi-agent technologies to study virus transmission patterns in complex urban systems,predict the development of epidemics under different interventions,and inform surveillance,early warning,prevention,and control policy decisions is of particular relevance to ensuring public health security,and the selected topic has both theoretical and applied value.Following is a list of the work in this thesis:(1)A decision-making method based on multi-agent technologies for epidemic surveillance and early warning to adjust prevention and control policies is proposed.An artificial social model of SARS-CoV-2 transmission,prevention,and control was designed and constructed on the Repast Simphony platform based on GIS using a combination of artificial social theory and multi-agent modeling methods.To address the shortcomings of existing models in the dynamic evolution of infectious disease transmission processes and the difficulty of reflecting human heterogeneity,humans in different states are abstracted as agents with self-control and autonomous decision-making capabilities,and the behavioral and interaction rules of agents are defined to portray the mobility,heterogeneity,contact patterns and dynamic interaction feedback mechanisms with the spatial environment of the population.By comparing the model with existing models and fitting the actual data,the average absolute error percentage of the constructed model is reduced by 3.45%,indicating that the model is widely applicable and can be used to judge the effectiveness of timely intervention measures,analyse the virus transmission status in complex urban systems and its change trend under different intervention measures,and provide new methods for urban epidemic prevention and control.(2)Using the six urban areas under the jurisdiction of Taiyuan as a case study,the model spatial environment was constructed based on the real geographical space and social environment as the model background,and the rules of state transition,flow and contact rules among five types of resident agents,rules for admission and treatment of patients by hospital agents,and rules for the decision-making behavior of government agents in formulating early warning and prevention and control policies.Based on this,multivariate simulation experiments were designed and conducted to evaluate the effectiveness of pharmacological interventions,non-pharmacological interventions,and combined intervention strategies to suppress the spread of the epidemic.The effectiveness of each intervention on the spread of the epidemic was verified,revealing the importance of implementing precision prevention and control in complex urban systems.(3)To address the limitations of single-scale simulation in reflecting regional differences and the accuracy of cross-scale multi-agent interaction and evolution,the mesoscale theory and method in the field of chemical industry are introduced in the design of multi-agent parallel simulation framework,and the model is divided into micro-scale,mesoscale and macro-scale.The framework of the mesoscale-based parallel simulation system is constructed by using the idea of hierarchical parallel computing according to the characteristics of cross-scale operation of the agent.The results show that by expanding the simulation scale of the model,the prediction accuracy and analysis efficiency of the multi-agent model can be improved. |
PDF Full Text Request