Bounded rationality in agent-based models: experiments with evolutionary programs