How much can Buffett influence the stock market? — A research of ACF

Traditional classifier system is considered as a multi-population GAs architecture which represents individual learning. In real world, people learn not only from individual experience but also from others. By introducing a concept called `social learning bonus´ an extended classifier system which mixed individual learning and social learning is implemented in an artificial stock market. The results suggested social learning leads to different market statistic property around a threshold, and a state like HREE may be realized endogenously.