This model incorporates all of the eight possible strategies of the well-known prisoners' dilemma with possible reciprocity based on memory of a single round. There is an initial strategy (C or D), a strategy that is employed following cooperation of a particular agent in the previous round (C or D), and a strategy that is employed following defection of a particular agent in the previous round (C or D). Thus, there are eight strategies: CDC, DDD, CDD, DCD, CCD, CCC, DCC, and DDC. The replication dynamic is by imitation. After each agent plays a round with each of eight neighbors (a Moore neighborhood) on an unbordered toroidal lattice, each agent sequentially adopts the strategy of the neighbor who has accumulated the highest adaptive score - a summation of payoffs from the matrix that can be specified in the simulation. Should there be a tie in these scores, one is chosen at random. If the focal agent's adaptive score is part of a tie, it keeps its own strategy if the status quo switch is on, and chooses one at random from all tied strategies, including its own, if the status quo switch is off. Imitation occurs with error at a rate specified by the mutation rate.

Pressing "Setup" initializes the taurus with random strategies. Pressing "Go" runs the simulation (Press "go" again to halt). Begin by trying the simulation with a mutation rate of zero. Then try the simulation with a variety of different mutation rates. What effect does the mutation rate have on the evolution of cooperation? Does it matter whether or not the status quo is maintained in ties?

© Copyright 2007, Gregory Todd Jones, All Rights Reserved.