Viscosity and dispersion in an evolutionary model of learning

Abstract

A two-population evolutionary model of learning is proposed where there is a conflict of interests between populations, and where interaction ranges from complete viscosity to complete dispersion. The long run stochastically stable states under complete viscosity match those of Kandori, Mailath and Rob (1993). With complete dispersion, the long run stochastically stable states match those of Hehenkamp (2001). With fluid interaction, there is a possibility of convention coexistence. Welfare properties of the long run stochastically stable states are examined using utilitarian and Rawlsian measures of welfare.