宏观异质性代理人模型及算法研究进展  

Research Progress of Macroscopic Heterogeneous Agent Model and Algorithm

作　　者：满百举 刘霞辉[1,2] 孙怡龙 Man Baiju;Liu Xiahui;Sun Yilong(School of Economics,University of Chinese Academy of Social Sciences,Beijing 102488,China;Institute of Economics,Chinese Academy of Social Sciences,Beijing 100836,China;Institute of International Economics and Politics,Shandong Academy of Social Sciences,Jinan 250002,China)

机构地区：[1]中国社会科学院大学经济学院,北京102488 [2]中国社会科学院经济研究所,北京100836 [3]山东社会科学院国际经济与政治研究所,济南250002

出　　处：《统计与决策》2023年第17期28-33,共6页Statistics & Decision

基　　金：国家社会科学基金青年项目(20CJL032)。

摘　　要：宏观理论的发展以及数据可得性的提高促进了异质性代理人模型的发展,系统梳理该模型的发展历程及求解算法十分必要。从算法角度讲,虽然离散时间异质性代理人模型的求解方法已有多种,但其求解逻辑无法规避状态变量的离散化迭代过程,因此难以解决“维数诅咒”问题。而连续时间异质性代理人模型的稳态则可由偏微分方程系统来表示,避免了在状态变量上的迭代过程,具有求解效率高、经济含义突出、适用性广、算法通用性强等特点。因此,连续时间异质性模型将会成为宏观模型的新主流。The development of macro theory and the improvement of data availability have promoted the development of heterogeneous agent model,and it is necessary to systematically sort out the development process and solution algorithm of the model.From the perspective of algorithm,although there are many methods to solve the discrete-time model,its solution logic cannot avoid the discretization iteration process of state variables,so it is difficult to solve the problem of“dimension curse”.The steady state of the continuous-time heterogeneous agent model can be represented by the partial differential equation system,which avoids the iterative process on the state variables,so it has the characteristics of high efficiency,prominent economic meaning,wide applicability and strong algorithm versatility.Therefore,the continuous-time macro heterogeneity model will become the new mainstream of macro models.

关 键 词：连续时间 异质性代理人模型 维数诅咒 微分方程系统 

分 类 号：F015[经济管理—政治经济学]