Professor of Shibaura Institute of Technology
Computational social science
especially, evolutionary simulation of a society
As for the formation of a state, since North and Olson, many predation theories have been presented, but have not considered an endogenous formation of a stable state. Then, developing a model of replicator equations based on Olson’s stationary bandit, we simulated the formation of a state. Generally, a replicator dynamic assumes a number of strategies (roles). Interactions occur among players with each role in every generation, and the above-average role in payoffs grows in frequency through generations. Players’ birth and death occur at the end of each generation, and it changes the role distribution in a society, which is described with a vector of roles’ frequencies. Since asymptotically stable points of a replicator dynamics can be regarded as a social structure (e.g., a state), we search for the stable points in the vector space and examine transitions among them. Particularly, the model consists of five roles: a peasant, roving bandit (bandit), peasant paying tribute (citizen), stationary bandit (guardian), and emigrant. Peasants engage in production and bandits rob peasants of their wealth. Guardians stop the harsh predation of citizens and citizens pays tribute to guardians. Emigrants can move to another location without predation or tribute, or back to the origin. We executed evolutionary simulation using the equations and found an asymptotically stable point consisting of guardians and citizens, that realizes a Pareto-efficient society. Based on the efficiency, following previous works, the point can be interpreted as the endogenous formation of a stable state. We also found that the stable point is more likely to appear when guardians adopt the comparatively low tribute rate and than without emigrant.