Advances in Artificial Life

Volume 2801 of the series Lecture Notes in Computer Science pp 138-145

Effects of Learning to Interact on the Evolution of Social Behavior of Agents in Continuous Predators-Prey Pursuit Problem

  • Ivan TanevAffiliated withATR Human Information Science Laboratories
  • , Katsunori ShimoharaAffiliated withATR Human Information Science LaboratoriesGraduate School of Informatics, Kyoto University

* Final gross prices may vary according to local VAT.

Get Access


We present the results of our work on the effect of learning to interact on the evolution of social behavior of agents situated in inherently cooperative environment. Using continuous predators-prey pursuit problem we verified our hypothesis that relatively complex social behavior may emerge from simple, implicit, locally defined, and therefore – robust and highly-scalable interactions between the predator agents. We argue that the ability of agents to learn to perform simple, atomic acts of implicit interaction facilitates the performance of evolution of more complex, social behavior. The empirical results show about two-fold decrease of computational effort of proposed strongly typed genetic programming (STGP), used as an algorithmic paradigm to evolve the social behavior of the agents, when STGP is combined with learning of agents to implicitly interact with each other.