How cognitive and environmental constraints influencethe reliability of simulated animats in groups

Abstract

Evolving in groups can either enhance or reduce an individuals task performance. Still, we know little about the factors underlying group performance, which may be reduced to three major dimensions: (a) the individuals ability to perform a task, (b) the dependency on environmental conditions, and (c) the perception of, and the reaction to, other group members. In our research, we investigated how these dimensions interrelate in simulated evolution experiments using adaptive agents equipped with Markov brains (\"animats\"). We evolved the animats to perform a spatial-navigation task under various evolutionary setups. The last generation of each evolution simulation was tested across modified conditions to evaluate and compare the animats reliability when faced with change. Moreover, the complexity of the evolved Markov brains was assessed based on measures of information integration. We found that, under the right conditions, specialized animats were as reliable as animats already evolved for the modified tasks, that interaction between animats was dependent on the environment and on the design of the animats, and that the task difficulty influenced the correlation between the performance of the animat and its brain complexity. Generally, our results suggest that the interrelation between the aforementioned dimensions is complex and their contribution to the groups task performance, reliability, and brain complexity varies, which points to further dependencies. Still, our study reveals that balancing the group size and individual cognitive abilities prevents over-specialization and can help to evolve better reliability under unknown environmental situations.\n\nAuthor SummaryThe ability to adapt to environmental changes is an essential attribute of organisms which have had evolutionary success. We designed a simulated evolution experiment to better understand the relevant features of such organisms and the conditions under which they evolve: First, we created diverse groups of cognitive systems by evolving simulated organisms (\"animats\") acting in groups on a spatial-navigation task. Second, we post-evolutionary tested the final evolved animats in new environments-not encountered before- in order to test their reliability when faced with change. Our results imply that the ability to generalize to environments with changing task demands can have complex dependencies on the cognitive design and sensor configuration of the organism itself, as well as its social or environmental conditions.