Decision-making Behavior and Welfare Distribution Mechanism of Economic Agents under the Constraint of Information Acquisition Cost

Abstract

This article aims to explore how the cost constraint of information acquisition reshapes the decision-making logic of economic subjects and affects the distribution pattern of social welfare. Aiming at the limitation that the traditional model ignores the cost of information search, a multi-agent reinforcement learning framework with heterogeneous agents is constructed, and the information budget is internalized as an immediate penalty in the decision-making process. By setting the low, medium and high cost parameters, a large-scale simulation experiment is carried out, and the market evolution characteristics under different constraint levels are analyzed. With the increase of information cost from 0.5 to 5.0, the average search frequency of subjects decreased significantly, the market bid-ask spread expanded to 1.8 times of the initial level, and the convergence rate of price error slowed down by 42%. In the high-cost situation, the Gini coefficient of wealth climbed above 0.65 in the later stage of the experiment, while in the low-cost group, it was only 0.32, indicating that the information barrier aggravated the marginalization of vulnerable groups. The high cost of information acquisition reduces market liquidity and pricing efficiency, and leads to uneven welfare distribution. Based on this, it is proposed to establish a public information platform and a dynamic compensation mechanism to lower the information threshold of specific groups through institutional innovation.