Population-Independent Pairwise Proportionally Imitative Dynamics for Multipopulation Games

Multipopulation games model interactions among many agents who have various roles. The imitative protocol describes when and how each agent changes her action in multipopulation games. Under existing imitative protocols, each agent imitates only agents who belong to the population that she belongs to. In other words, it is assumed that each agent knows the populations that the other agents belong to. However, the agents cannot always know the populations that the other agents belong to. In multipopulation games that consist of agents whose senses of values are different from one another, we consider situations where agents cannot know populations that the other agents belong to. We propose a population-independent imitative protocol, and derive the mean dynamics that is generated by the protocol.