Laboratory expriment and evolutionary competition in lowest unique integer games
This study examines what kind of behavioral models are successful in small-sized lowest unique integer games by laboratory experiment and evolutionary competition. For this purpose, we use the behavioral data from out past laboratory experiment, estimate the parameters of typical learning models for each subject and then pursue computational experiments. The main but preliminary findings are in the following: First, sticking behaviors such as Level-1 and Level-2 perform well both in evolutionary competition and in laboratory experiment. Second, as the generations goes by, the frequency of average changes decreases while the game efficiency differs from the game setup, namely improved in three-person LUIGs but not in four-person LUIGs.