Altruism in the Wilderness
| So, what happens when a mixed community of altruists and defectors
lives on the edge of an increasingly hostile environment? In this simulator, agents occupy pacthes alone or with other agents. They play prisoner's dilemma with other agents in the same patch. Patch extinction rates range from zero in the upper left hand corner to 3% along the right and bottom edges. The initial population of 40 cooperators and 40 defectors starts in the benign portion of the grid. Agents have a 2% probability of moving each cycle. PD payoffs use a 1:4 cost:benefit ratio.
Resource grows at a rate of 5 units per cycle, with an upper limit of 50 units. Agents' "metabolisms" consume one unit of resource per cycle, so patches have a carrying capacity of 5 agents in the absence of play. Agents breed if they accumulate 150 units of resource, either from play or foraging. The grid is not wrapped. The color of the agents gives their resource level, blue for low, red for high. White bars are cooperators, black bars are defectors. You can get some information about the population in a patch by clicking on it. In the absence of extinction, cooperators are able to migrate out and survive for a while, but eventually defectors will invade and drive them to extinction. What happens if exinction rates increase as you go farther away from the population source? How long can altruists survive in the wilderness? Check it out: FrontExtGame.html The Emergence of Cooperative Behavior Can altruists invade defectors at the edge of survivabilty? Start the simulator with
all defectors, and a 0.005 mutation rate in reproduction. (Turn off drawing to speed things up for a while.) Evolving Discriminating Mechanisms Finally, I have been speculating that these kinds of edges of evolvability provide the conditions under which discriminating mechanisms can evolve. The simplest and most popular way of thinking about discriminating mechanisms is the strategy Tit-For-Tit (TFT). In this scenario, agents play a repeated prisoner's dilemma, and are able to condition their move on the opponent's last move. TFT starts out by cooperating, and thereafter does whatever the opponent did on the last move. A scenario in which TFT mutates from Always-Defect, with a Benefit:Cost ratio of 2:1, and three repetitions can be simulated with only a minor modification of the payoff matrix.
Start-er-up with the payoffs changed: FrontExtGame3.html |