| On 30 January 2020,WHO issued the highest alert level and declared novel coronavirus（Coronavirus disease 2019（COVID-19））a "public health emergency of international concern",and then declared it a "global pandemic" on March 11,2020.Due to the lack of protective vaccine and effective therapeutic drug intervention methods,coupled with a large number of asymptomatic cases,presymptomatic transmission and strong infectivity,COVID-19 has spread rapidly around the world,bringing great challenges and significant impacts to the production and life of countries around the world.In order to deal with the epidemic,relevant researchers have actively engaged in research work,and carried out multiperspective studies on the source and transmission of the virus,the prevention and diagnosis of the disease,the care and treatment of patients,the impact and countermeasures of the epidemic,the methods and measures of epidemic prevention and control,and the effects of the epidemic.Current researches are mostly carried out in the fields of epidemiology and medicine,and their methods are mostly based on statistical data.Statistical analysis and mathematical modeling methods are adopted to conduct prediction analysis from the macro level.For global pandemic infectious diseases,the attribute characteristics of social individuals and their contacts are the key factors for the development of the epidemic.In the existing studies,individuals are taken as the basis and simulation models are used to depict the transmission of infectious diseases from the bottom up,and whether the prevention and control measures are effective or not,and there are few studies on the effects of different prevention and control measures.Therefore,this paper takes COVID-19 as an example to study the construction of artificial population in large cities and its application in epidemic prevention and control,with a view to providing high-precision basic population data sets and simulation research methods for epidemic prevention and control.In this paper,firstly,the literature and basic theories and methods of artificial population and epidemic situation are reviewed.Then,Wuhan is taken as the research object to explore different micro population data sources,different sampling methods,and effective methods for the construction of artificial population in large cities.On this basis,based on the constructed high precision artificial population in Wuhan,a multi-agent epidemic transmission simulation model was established to dynamically depict the epidemic transmission in the city,and the effectiveness of the prevention and control measures was studied.Finally,corresponding policy suggestions are put forward according to the research conclusions.The main conclusions are as follows:（1）The quality of urban artificial population generation mainly depends on the quality of micro population data sources.The more comprehensive the attribute information of micro population data is,the higher the stability and precision of the synthetic artificial population data set will be.The larger the sample size of micro data,the higher the precision of the synthetic artificial population data set will be.After testing and comparison,it is found that due to the detailed information of population attribute and large sample size,the synthetic artificial population data set is more stable and more accurate than the data set generated by "World Micropopulation Database"（IPUMS）.（2）Simple random sampling method（SRS）,moment matching algorithm（MM）and iterative proportional fitting algorithm（IPF）are the main methods for artificial population synthesis in the SPEW system,and the choice of synthesis method will affect the accuracy of synthetic artificial population dataset.Compared with the SRS method and the MM algorithm,the IPF algorithm has a higher and more stable accuracy,and the accuracy remains above 85%.Specifically,in the 2010 artificial population data set,the distribution of population attributes under the SRS method is relatively concentrated,which cannot reflect the difference of demographic characteristics in different areas of the city.MM algorithm is constrained by the average size of real household.Although the size of household is more accurate than other algorithms in the synthetic data,the accuracy of MM algorithm in the gender and age structure of population is far less than that of IPF algorithm.Under the joint constraint of household size distribution and household head race distribution,the artificial population attribute distribution synthesized by IPF algorithm is closer to the real demographic characteristics at both the whole and regional levels of Wuhan.Therefore,when synthesizing the artificial population of big cities in China,it is more appropriate to adopt the IPF algorithm if the SPEW system is adopted.（3）The artificial population data set generated based on the SPEW system,because its attributes are really close to the urban demographic characteristics,ensures the individual heterogeneity and group linkage of urban disease transmission during modeling,and has good application value in the simulation research of epidemic prevention and control.The empirical analysis found that the simulation of COVID-19 epidemic transmission based on the artificial population data with real demographic characteristics was more accurate than the traditional numerical simulation in carrying out the simulation research on prevention and control policies. |
PDF Full Text Request