Knowledge Diffusion Analysis as an Evolutionary Advantage in a 2D Grid Virtual Environment

Abstract

A prototype is developed to simulate the diffusion of knowledge among intelligent agents to evaluate the hypothesis that the agents obtain an evolutionary advantage when they share information from their environment. For this purpose, the emotional model proposed by (Mehrabian, 1997) is integrated to the agents together with the adaptation of the genetic algorithms, and the prisoner’s dilemma to simulate the interaction between them and the decision-making in front of the events. During the test analysis phase, a high correlation is observed between the agents that manage to live in the last generations, and the amount of knowledge they obtain about the environment. There is also a more significant presence of the emotions of anger and happiness over fear and hope throughout the execution of scenarios, as well as very optimistic and very pessimistic personalities among the surviving agents. On the other hand, the integration of Evolutionary and Affective Computing together with game theory allows the generation of emerging behaviors within the environment. In addition, the models implemented were appropriate to the paper’s objectives.