What is Polyworld?

Polyworld is computer programmer Larry Yaegar’s Artificial Life (also known as Alife and Theoretical Biology) computer simulation, or computational ecology which seeks to mirror real life’s ecological and anthropological development. The environment in which Polyworld’s inhabitants operate in is available for installation across a number of platforms (Mac OS X, Linux) and involves a 2-dimensional plane that is rendered using the QT graphics toolkit and OpenGL.

The entities that populate the Polyworld environment act in accordance with their underpinning neural network and demonstrate the whole gamut of human action and interaction; that is, they can eat, drink, reproduce, battle with and kill each other. In short, the simulated organisms either develop survival strategies and prosper or don’t and die.

The behavior of the individual simulated organisms that inhabit Polyworld is dictated by each organism’s behavioral suite and the underlying neural network that serves as the organism’s brain. The behavioral suite is the name given to the organism’s capacity to move, turn, attack, eat, mate and light. The neural network or brain architecture is the name given to the types and arrangements of the neural clusters that go to make the organism’s brain. These neural clusters are made up of neurons that either excite (excitatory neurons) or inhibit (inhibitory neurons) the entity, and in effect, dictate the personality and ultimately the survival of the various Polyworld beings.

The synaptic efficiency of the Polyworld’s various entities operates according to the laws of Hebbian learning. This allows a Polyworld entity - by repeatedly firing the same neurons – to develop and learn from its past experience by strengthening its neural connections and developing the neurological topography that has proven effective in promoting its well being and survival in the past.

Both the organism’s physiology and its brain are encoded in its genetic make-up, and therefore the Polyworld environment is populated and will continue to be populated by the various evolved generations of a particular original forbear. This has meant that clans or species have evolved who display their species-particular behavior and strategies for survival. These strategies have manifested themselves in a number of sophisticated traits of human behavior that include escaping, avoiding, foraging, swarming and flocking. Polyworld has also been updated to support parallel "driven" vs. "passive" simulations in an effort to discern the nature and existence of an evolutionary "arrow of complexity."