Emergence of altruistic behavior in heterogeneous populations of artificial agents by evolution of kin discrimination mechanism

The emergence of altruistic behaviors in heterogeneous populations of autonomous robots, especially in signaling tasks, has proven to be a difficult problem to solve. However signaling and altruistic behaviors are present throughout the tree of life. Specially giving that, signaling behaviors seem to have evolved multiple times whenever there is a channel to emit a signal and one to receive it. In this work, this problem is addressed, using evolutionary algorithms, and modeling phenomena such as kin selection and kin discrimination in a biologically plausible way. We also used self-organizing maps to analyze the behavior of these populations during the evolutionary process, within the solution space. We believe that this approach can shed light on the predictive power of the Hamilton rule, the importance of kin selection in the evolution of altruistic behaviors, and how self-organizing maps can allow us to observe the different solutions in which the evolutionary algorithm converges through time.