Polyworld
An Artificial Life System and Computational Ecology


Of particular note to visitors and developers:

 

A few modern, high-quality Polyworld Movies have been added to the site.

 

The initial setup and build process for Polyworld has been somewhat streamlined and documented with careful step-by-step instructions.

 

Recent news (in reverse chronological order):

 

A paper has been accepted to the ALife XII conference, discussing our efforts to relate network structure to network function, that uses standard network science metrics (and defines a couple new ones) to demonstrate robust evolutionary selection for small-world network topologies, comparing multiple driven runs to passive null-model runs, and to highlight a convergence between physical constraints on network structure (minimizing wiring length and volume) and evolutionary pressures for network functionality (as measured by our usual information theoretic measure--simplified "TSE complexity"--of neural dynamics).  Humans and other organisms of relatively high intelligence are probably all beneficiaries of this fortuitous convergence between physical and functional constraints on neural network topologies.  The paper is available below.

 

A paper primarily by some collaborators at CSIRO in Sydney, Australia, documenting an evolutionary trend towards "small world" networks in parallel with the trend towards greater complexity in Polyworld, has been accepted for an oral presentation at ECAL 2009.  A preprint is available below.

 

An article expanding on the driven vs. passive results from Polyworld, their relation to behavioral adaptation, and the distinction between complexity derived from natural selection vs. other mechanisms (genetic algorithms and random walks) is to appear in the HFSP (Human Frontier Science Program) Journal soon.  A preprint is available below.

 

Polyworld has been updated to support parallel "driven" vs. "passive" (using McShea's terminology) simulations in order to investigate the existence and nature of any evolutionary "arrow of complexity".  It also supports the use of neural complexity (after Sporns's work) as an evolutionary fitness function.  A paper discussing these results was presented at the ALife XI Conference, and may be found below.

 

Two Polyworld-based papers were presented at the ALife X conference.  Links to the papers may be found below.  One paper addresses a common behavioral ecology/evolutionary biology issue—how agents distribute themselves given limited, patchy resources—while the other paper applies an information-theoretic measure of complexity to the neural functioning of agents in Polyworld, providing both a long sought quantitative assessment of complexity in such systems and evidence for a statistically significant increase in complexity over evolutionary timescales.

 

Polyworld "lives" again!  As a result of the research freedom offered by my faculty position in the School of Informatics at Indiana University, I have, with the extensive help of Gene Ragan and Nicolas Zinovieff, revived and modernized Polyworld.  It has been ported to run atop Qt, from Trolltech, and OpenGL, so it should be fully cross-platform (Mac OS X, Windows, and Linux) and should run on any modern personal computer or workstation, although only Mac OS X and Linux versions have been successfully deployed so far.  The app has about 90% of its old functionality, and the remaining 10% is mostly just user interface stuff—the core simulation engine is fully functional, I believe. This new version of the source code was released via SourceForge on December 25, 2004 and may be accessed at http://sourceforge.net/projects/polyworld/.  Substantial new capabilities have since been added.

Polyworld background and details:

Polyworld is a computational ecology that I developed to explore issues in Artificial Life. Simulated organisms reproduce sexually, fight and kill and eat each other, eat the food that grows throughout the world, and either develop successful strategies for survival or die. An organism's entire behavioral suite (move, turn, attack, eat, mate, light) is controlled by its neural network "brain". Each brain's architecture--it's neural wiring diagram--is determined from its genetic code, in terms of number, size, and composition of neural clusters (excitatory and inhibitory neurons) and the types of connections between those clusters (connection density and topological mapping). Synaptic efficacy is modulated via Hebbian learning, so, in principle, the organisms have the ability to learn during the course of their lifetimes. The organisms perceive their world through a sense of vision, provided by a computer graphic rendering of the world from each organism's point of view. The organisms' physiologies are also encoded genetically, so both brain and body, and thus all components of behavior, evolve over multiple generations. A variety of "species", with varying individual and group survival strategies have emerged in various simulations, displaying such complex ethological behaviors as swarming/flocking, foraging, and attack avoidance.

For further information, please feel free to take a look at (in reverse chronological order, except for the original Polyworld paper that appears first):

Yaeger, L. S. 1994. Computational Genetics, Physiology, Metabolism, Neural Systems, Learning, Vision, and Behavior or PolyWorld: Life in a New Context. Langton, C. ed. Proceedings of the Artificial Life III Conference. 263-298. Addison-Wesley.

Yaeger, L., Sporns, O., Williams, S., Shuai, X. and Dougherty, S. 2010. Evolutionary Selection of Network Structure and Function. In Artificial Life XII: Proceedings of the Twelfth International Conference on the Simulation and Synthesis of Living Systems. (Accepted)

Lizier, J.T., Piraveenan, M., Pradhana, D., Prokopenko, M., Yaeger, L.S. Functional and Structural Topologies in Evolved Neural Networks. ECAL 2009 (preprint).

Yaeger, L. S. 2009. How Evolution Guides Complexity. HFSP Journal (preprint).

Yaeger, L. S., Griffith, V., and Sporns, O. 2008 (accepted). Passive and Driven Trends in the Evolution of Complexity. In Bullock, S. et al. eds. Artificial Life XI: Proceedings of the Eleventh International Conference on the Simulation and Synthesis of Living Systems. MIT Press. Cambridge, MA.

Griffith, V. and Yaeger, L.S. 2006. Ideal Free Distribution in Agents with Evolved Neural Architectures, in Rocha, L. et al. eds. Artificial Life X: Proceedings of the Tenth International Conference on the Simulation and Synthesis of Living Systems. 372-378. MIT Press. Cambridge, MA.

Yaeger, L. S. and Sporns, O. 2006. Evolution of Neural Structure and Complexity in a Computational Ecology, in Rocha, L. et al. eds. Artificial Life X: Proceedings of the Tenth International Conference on the Simulation and Synthesis of Living Systems. 330-336. MIT Press. Cambridge, MA.

Yaeger, L. S. and Sporns, O. 2006. Evolution of Neural Complexity.

The README.txt file has some info on the subject, but complete instructions for installing and building Polyworld are here. There is also a short list of outstanding to-do items in the TODO.txt file, divided into simple code restoration/maintenance tasks and basic research directions.

 

I've had an ongoing interest in the subject of Artificial Life since 1987 (and in related subjects, such as how the mind works, since even earlier), and have placed some of my notes and conference reports in the ALife section on this server.