A study of heterogeneity in a stock market simulator based on a model of agents that learn from experience in a market with multiple stocks

Heterogeneity of agents in stock markets has been the subject of many works lately. Many interesting results have been raised from these works, such as that the very large volume of trade should be a result of the heterogeneity of agents, more precisely, of the existence of chartists and fundamentalists in the same market. Another interesting result is that, in some simulations, heterogeneity reduces temporarily, with a large number of fundamentalists behaving as chartists during certain periods, to return to behave as fundamentalists after some time. The objective of this work is to replicate those results in a stock market simulator based on a model of agents that learn from experience in a market with multiple stocks, and verify how the learning process of the agents interferes in the distribution of agents’ profiles. In addition, the fact that the market have multiple stocks has allowed us to study the correlation between an index based on a basket of stocks and such distribution.