Research On Agent Model For The Third Distribution Of Wealth

At present,there is a significant gap in the distribution of wealth among Chinese residents.In order to achieve the model of common prosperity,the country has continuously introduced relevant policies for the three distribution of wealth.However,in the current research based on the triple distribution of wealth,domestic experts and scholars mostly focus on the theoretical research of the existing policies.Compared with the multi-agent technology,which can carry out forward-looking analysis of the forthcoming relevant policies,the theoretical research needs to be further improved and explored.Based on the theory of Multi-Agent System(MAS),this thesis uses multiagent technology to simulate the evolution process of the economic field,and analyzes the impact factors of the third distribution of wealth,which can simulate the impact of the simulation policy,and is of great significance to the realization of the common prosperity model.First of all,in order to simulate the evolution process of the complex adaptive system of the third distribution of wealth in the economic field,the reinforcement learning algorithm is used to train the agent,so that the agent has the self-organization and adaptive ability,and has the characteristics of the real economic society.Secondly,through the analysis of the existing wealth distribution model,the factors such as the level of knowledge,the degree of preference,the transaction time and the saving rate of the agent are mainly studied.Based on the theory of rarefied gas dynamics,the joint density image of the level of knowledge and wealth distribution is obtained;Using the utility function of Constant Elasticity of Substitution(CES),this paper studies the impact of different preferences of traders on commodity prices and wealth distribution;Based on the discrete time domain wealth distribution model,the influence of transaction time on wealth distribution is studied.Finally,these factors are defined in the agent model,and the Net Logo simulation platform is used to carry out simulation tests to explore the impact of model parameters on the evaluation index-Gini coefficient,and give the policy recommendations for the third distribution of wealth.This paper tries to integrate disciplines,applying artificial intelligence technology to the research of wealth distribution three times in the economic field,giving full play to the advantages of the integration of disciplines,and greatly promoting the ability of academic research and solving practical problems.The main research achievements of this paper are as follows:1.Construct the Agent model of wealth three times distribution.According to the research and analysis,Agent types and attributes are defined,Agent behavior activities are analyzed,behavioral decision-making mechanism is developed,evaluation functions are selected and agent simulation model framework is constructed,and then influence factors of wealth distribution are analyzed,and relevant policy suggestions are put forward.2.Build a model library.By defining Agent attributes,behavior rules,decision mechanism and model framework,and finally running on Net Logo simulation platform,a model library function is provided for researchers to facilitate research and analysis in different environments.