| Our country is a developing country with a large population and relatively backward public health conditions, thus emergent biological events have been an important problem threatening the safety of the people and social stability. The outbreak of SARS in 2003 has made people turn pale at the mention of its name even now. As one of the hardest hit areas of "SARS", despite that our government mobilized the whole power of the country and all the people united to fight the epidemic, "SARS" still caused a serious loss of life and property in China. The outbreak also exposed many deficiencies of our country’s ability to deal with unexpected biological events and maintain national biosafety.At this stage, China is still facing many kinds of biological threats. In recent years anthrax attacks and ricin letters occurred within the United States, indicating the terrorists have begun to use biological agents in terrorist activities. Biological terrorist attacks is a potential factor threatening the social stability of our country. Despite the repeated calling from the international community for strictly prohibiting the use or research of biological weapons, some countries are still doing research on biological weapons. As the primary victim of biological weapons in the history, China still needs to be vigilant of the potential threat of biological weapons. As a main kind of biological security events threatening our country, emerging infectious diseases in recent years, such as the outbreak of SARS, H5N1 bird flu, H1N1 and H7N9 bird flu, all seriously impact our life safety and social development. Although the 2014 Ebola outbreak in West Africa did not cause much impact on China, it also underlined that the threat of emerging infectious diseases emergencies will always exist. In addition, the leak problem of biological laboratories all over the world may also be a major factors causing biological events. Therefore, the biosafety capacity of our country should be strengthened.Recently, with the development of high-performance computing and big data processing techniques, simulation of various biological events has become a hot research field of bio-safety, and also be the foundation of national biosafety capacity. Foreign research institutions have established and put into use several bio-event simulation systems, make great contributions for the optimization of infectious disease control strategies and biological events contingency plans. However, overall domestic simulation research about biological events is still at an early stage. There is a big gap between domestic research and advanced foreign research from the collection of raw data and the set-up of virtual environment to the building of the spread model of infectious diseases and diffusion model of aerosol.Generating synthetic populations is an important foundation work for the build of most bio-event simulation systems. Synthetic populations are the virtual people created in computers while keeping statistically consistent with the real people, rather than one-by-one copy of the real people. As an important basis for agent-based modelling, synthetic populations are widely used in the simulation of infectious diseases spread, traffic flow, battlefield dynamics and policy results. Depending on the purpose of research, required attribute of synthetic populations also varies. Synthetic populations for bio-event simulation in general require fitting individual age, gender, household kind, household position, and so on. The main purpose of this paper is to study the generation technology of synthetic populations to support biological event simulations.In existing studies, traditional methods for generating synthetic population is Synthetic Reconstruction and Combinatorial Optimization based on sample data. Both can generate synthetic populations conform to demographic profile, but they need discrete sample data, which limits the application of these two methods. Especially for China, due to the lack of available demographic sample data, it is impossible to directly use the traditional methods to generate synthetic populations of our country. In recent years, some researchers have proposed new generation methods needless of sample data, but most of these techniques are proposed for the West household. The normal structure of households in Europe or America is relatively simple. Contrarily, there are many complex multigenerational households in China. Therefore, in order to generate accurate synthetic populations of our country, this study focused on solving the generation of multigenerational households.To solve these problems, this paper carried out the following three research tasks:(1) Established a set of essential data needed to generate synthetic populations.The essential data includes two categories, namely, demographic data and geographic information data. Demographic data used here is primarily from the sixth China census data in 2010, which includes the basic statistical properties of people. In addition, we also used the part data of the Chinese General Social Survey as a supplement to the census data. GIS data used here includes a border map of Beijing towns and gridded Beijing population distribution data, which is generated by the authoritative research institutions and can be obtained through public sources. Additionally, this article also uses Land Scan global population distribution data. These data provide the basic information needed to generate synthetic populations. Demographic data includes statistical distribution of a variety of demographic attributes, which can be used to generate the social attributes of synthetic populations. Geographic information data is mainly used to generate the spatial distribution of synthetic population.(2) Proposed a new generation method of synthetic populations, which uses the basic household unit to assign members for multi-generational households.Firstly, generate a basic set of virtual individuals and virtual households based on collected demographic data. Attributes generated include individual age, gender, household type, towns ID, as well as number of household members, number of member generations, town ID of each household. Then define the marital relationship and mother-child relationships between individuals according to the statistic data. B Build the basic household units which consist of relatives of one for two generations based on these two kind of individual relationship. Finally, according to the number of members and the number of generations of the virtual households, assign the basic household units into different virtual households to complete the assignment process of household members. Moreover, this paper generates virtual group quarter. Assign locations based on geographic information for the virtual households and group quarters to get the geographic distributions of the population.(3) Verify the validity of the new synthetic population generation method.Beijing synthetic populations generated by the above method contain two levels of data, individual data and household data. By analyzing the relationship between husband and wife and relationship between mother and children of the virtual individuals, prove that the method can more reasonably simulate the distribution of individual properties. By comparing with the results of randomly assignment, verify the reasonableness of household members. In addition, by comparing with Landscan population distribution data, evaluate the results of population geographic distribution defined based on the gridded data. The results showed that the present study can solve the problem of constructing synthetic populations.Research of generation method for fine synthetic populations is a major hotspot of social research. Compared to the word abroad, two major issues should be solved to generate synthetic populations of China. The first is the lack of publicly available data sample of the population, and the second is complex multi-generational household membership structure. This paper put forward an innovative method to build synthetic populations based on the basic household unit. This method does not require the use of population sample data, and can generate more reasonable member structure of the households. Synthetic populations generated by this method not only can support the simulation of biological events, it can also be used to simulate a variety of research areas such as transportation, policy analysis and so on. |
PDF Full Text Request