摘要
有别于信息物理系统(CPS),信息物理社会系统(CPSS)必须考虑参与者理性及非理性行为的影响。单独应用理论分析、模型仿真、实验经济学或数据驱动方法,都难以支撑CPSS的研究。混合仿真方法整合上述研究范式的长处,融合统计分析、因果分析与行为分析,可有效地支撑CPSS的研究。为此,除了CPS研究中涉及的数学模型外,还必须构建多代理模型来反映理性行为的高维不确定性。在反复仿真的过程中就只需要少数参与者来反映非理性行为的影响,大大降低了对合格参与者规模的需求。文中结合电动汽车用户的购置意愿、出行意愿、碳市场交易决策行为方面的研究案例,揭示了混合仿真方法的实践可行性。
Different from the cyber-physical system(CPS),the cyber-physical-social system(CPSS) must consider the influence of rational and irrational behaviors of participants.It is difficult to support the research on CPSS by using theoretical analysis,model simulation,experimental economics or data-driven method separately.The hybrid simulation method gathers the advantages of the above research paradigms and integrates statistical analysis,causal analysis and behavior analysis,which can effectively support the research on CPSS.To this end,in addition to the mathematical models involved in the research on CPS,a multi-agent model must be constructed to reflect the high-dimensional uncertainty of rational behaviors.As a result,only a few human participants are needed to reflect the impact of irrational behaviors in the process of repeated simulations,which greatly reduces the demand for the scale of qualified participants.Finally,the practical feasibility of hybrid simulation method is revealed combined with the case studies of the electric vehicle users' purchase willingness,travel willingness and decision-making behavior of carbon market transactions.
作者
薛禹胜
吴巨爱
谢东亮
黄杰
蔡斌
XUE Yusheng;WU Juai;XIE Dongliang;HUANG Jie;CAI Bin(State Grid Electric Power Research Institute(NARI Group Corporation),Nanjing 211106,China;College of Automation&College of Artificial Intelligence,Nanjing University of Postsand Telecommunications,Nanjing 210023,China)
出处
《电力系统自动化》
EI
CSCD
北大核心
2023年第16期1-9,共9页
Automation of Electric Power Systems
基金
南瑞集团有限公司科技项目(信息-物理-社会元素的交互及协调技术,GF-GFWD-210338)。
关键词
信息物理社会系统
理性行为
非理性行为
多维联合概率
多代理建模
混合仿真
cyber-physical-social system(CPSS)
rational behavior
irrational behavior
high-dimensional joint probability
multi-agent modeling
hybrid simulation
