On artificial adaptive agents models of stock markets

The complex behavior of share prices in a stock market is studied under a modeling technique of artificial adaptive agents. Individual agents who are active in the market are identified and represented by mathematical functions. Share price is then determined by an arithmetic sum of these functions. Iterations of the models produce a time series of share prices, which exhibits nonlinearities similar to those found in real-world stock markets. Several experiments are reported in this paper. The wealth held by an agent at the beginning of the experiment and the method by which the agent adapts himself to market trends are shown to be important to the success of the agent.