Coevolution of Competitive and Cooperative Agent Behavior
Since 2009
Realistic and complex behaviors for artificial agents can be inspired by real-life predators and prey. The predator-prey domain can be used to study the evolution of both competitive and cooperative behaviors because the agents cooperate within teams and compete across teams. If teams of artificial predators and prey are coevolved, an arms race emerges. That is, if the predator team is more successful in the first few generations, the prey team quickly evolves to out-survive the predators in the following generations, and so on. A neuroevolutionary architecture called Multi-Component ESP promotes this arms race.

More recently, we have been collaborating with biologists (Dr. Kay Holekamp and Marc Wiseman) at Michigan State University who study spotted hyena behavior. Hyenas sometimes cooperate to capture large, difficult-to-catch prey like zebras, but mostly prefer to hunt alone, feeding on smaller prey. Experiments were designed to study some factors that affect the evolution of cooperation in predators during hunting:
  • shared or individual reward upon prey capture
  • communication among predators
  • net return from catching a particular type of prey
Hyenas have many ways of communicating (vocal, visual, tactile) and the next goal is to simulate the evolution of these various kinds of communication in artificial predators.

This research is supported in part through the BEACON Science and Technology Center for Evolution in Action, funded by NSF DBI-0939454.

Risto Miikkulainen Faculty risto [at] cs utexas edu
Padmini Rajagopalan Ph.D. Student padmini [at] cs utexas edu
Aditya Rawal Ph.D. Student aditya [at] cs utexas edu
Kay E. Holekamp Collaborator holekamp [at] msu edu
Evolution of a Communication Code in Cooperative Tasks 2012
Aditya Rawal, Padmini Rajagopalan, Risto Miikkulainen and Kay Holekamp, In Artificial Life (13th International Conference on the Synthesis and Simulation of Living Systems), East Lansing, Michigan, USA 2012.
The Role of Reward Structure, Coordination Mechanism and Net Return in the Evolution of Cooperation 2011
Padmini Rajagopalan, Aditya Rawal, Risto Miikkulainen, Marc A. Wiseman and Kay E. Holekamp, In Proceedings of the IEEE Conference on Computational Intelligence and Games (CIG 2011), Seoul, South Korea 2011.
Constructing Competitive and Cooperative Agent Behavior Using Coevolution 2010
Aditya Rawal, Padmini Rajagopalan and Risto Miikkulainen, In IEEE Conference on Computational Intelligence and Games (CIG 2010), Copenhagen, Denmark, August 2010.