Agents and Patches: The First Generation Simulator      History

Typically, investigations of the dynamics of strategy proliferation in "game-playing" agents are explored (1) at the population level (see Evolution and Information), (2) via "tournaments" of individual agents in small populations, where each player has an equal chance of playing each other player (Axelrod 1978), or (3) in celluar automata (see Cellular Automaton PD) where the agents have fixed postitions and change strategies based on how their neighbors are doing.

The E.A.M.E. Agent-and-Patch simulator is designed to introduce a number of additional elements of realism into this investigation, each of which can be turned off to study the effects of the others in isolation.

1. Agents are always someplace, namely, on a patch. Patches can grow a generic resource, and agents can feed on that resource.
2. Agents can get information about their patch, and choose behavior on the basis of that.
3. Agents can move to adjacent patches. This movement may or may not incur a cost.
4. Agents only play against other agents on the same patch - they play locally.
5. Agents can have innate "strategies", which differ.
6. Agents can have "tags" that are recognizable by others.
7. Agents can reproduce if they gather enough resource. Their offspring will be on the same patch and carry the same strategies and tags.

Lots of variable can be controlled via inputting "codes" via the "control" panel. Non-default values of those codes are recorded in the "session" panel. Full records of session parameters can be cut and pasted into a text editor (like "notepad"), and individual strings of codes (including the random seed) can be pasted into the "control" panel. See the instructions for more information.

Try hitting the Code List link on the left panel in order to get help with controlling the simulator.