The networked structure of investor relationship has been documented by many researches. This paper proposes a dynamic model based on investor network, and investigate the impacts of investor interactions on market stability. Two learning patterns are taken into account. The herd behavior is characterized with the imitation and diffusion of trading strategies in investor network. The behavioral equilibria and their stability conditions are studied with mathematical analysis. As the findings show, herd behavior can affect trader structure as well as market stability. Although it is a potential source of bubbles, there exist some cases where herd behavior is beneficial to market stability. The impacts of network structure are closely related with the imitation mechanism. If investors mainly care about the absolute number of neighbors who adopt different strategies, the network effect could weaken market stability. To guarantee the stability conditions, the variance of degree distribution should fit with other parameters. A high variance of degree distribution might lead to market instability.

许多研究都证明了投资者关系的网络结构。本文提出了一个基于投资者网络的动态模型，并研究了投资者互动对市场稳定性的影响。考虑了两种学习模式。从众行为的特点是交易策略在投资者网络中的模仿和扩散。通过数学分析研究了行为平衡及其稳定性条件。正如调查结果所示，从众行为会影响交易者结构和市场稳定性。尽管它是潜在的泡沫来源，但在某些情况下羊群行为有利于市场稳定。网络结构的影响与模仿机制密切相关。如果投资者主要关心采取不同策略的邻居的绝对数量，网络效应可能会削弱市场稳定性。为保证稳定性条件，度分布的方差应与其他参数相适应。度分布的高方差可能导致市场不稳定。