基于遗传算法的人工人口生成与应用研究  

Research on Artificial Population Generation and Application Based on Genetic Algorithm

作　　者：张红历[1] 邓井双 Zhang Hongli;Deng Jingshuang(School of Management Science and Engineering,Southwestern University of Finance and Economics,Chengdu 611130,China)

机构地区：[1]西南财经大学管理科学与工程学院,四川成都611130

出　　处：《系统仿真学报》2023年第9期1965-1974,共10页Journal of System Simulation

基　　金：国家社会科学基金(17BTJ006)。

摘　　要：高精度微观人口数据是疾病传播、交通出行、应急事件等仿真系统的关键基础数据之一,现实中多采用计算机生成人工人口进行模拟。出于计算效率和生成步骤规范化考虑,目前人工人口合成多采用迭代比例拟合法,但是它对基础数据要求严格,存在零单元和数据代表性偏差问题且无法同时保证个体和家庭层面的拟合。为克服这一不足,提出使用模拟退火算法生成初始解和精英选择策略的改进遗传算法合成人工人口,探究遗传算法的合成步骤与参数设置,并以生成的成都市人工人口为基础,进行生物气溶胶扩散仿真实验。实验表明,采用遗传算法合成人工人口的精度优于传统方法,对基础数据要求低,这一数据集可为仿真研究提供精细尺度的人口多属性信息。High-precision micro-population data are one of the key basic data for simulation systems such as disease spread,traffic travel,and emergency events.In reality,computer-generated artificial populations are often used for simulation.Due to computational efficiency and standardization of generation steps,the iterative proportional fitting method is currently used for artificial population synthesis.However,it has strict requirements on basic data and faces zero-unit and data representational deviation problems,and it fails to guarantee the fitting at the individual and family levels at the same time.In order to overcome this deficiency,an improved genetic algorithm using a simulated annealing algorithm to generate an initial solution and elitist selection strategy is proposed to synthesize artificial populations and explore the synthetic steps and parameter setting of the genetic algorithm.Based on the generated artificial population in Chengdu,a simulation experiment of biological aerosol diffusion is carried out.Experiments show that the accuracy of artificial populations generated by the genetic algorithm is better than that generated by traditional methods and has low requirements for basic data.This data set can provide fine-scale multi-attribute population information for simulation research.

关 键 词：人工人口 遗传算法 模拟退火算法 精英选择 气溶胶扩散 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]