Games

Here is an example of the Hawk Dove game:

from games.game import Game

class HawkDove(Game):
    DEFAULT_PARAMS = dict(v=30, c=60)
    STRATEGY_LABELS = ('Hawk', 'Dove')
    PLAYER_LABELS = ('Player 1', 'Player 2')
    EQUILIBRIA_LABELS = ('All_Hawk', 'All_Dove', 'Hawk_Dove', 'Dove_Hawk')

    def __init__(self, v, c, equilibrium_tolerance=0.1):
        payoff_matrix = (( (v - c) / 2.0, v),
                         (0, v / 2.0))

        player_dist = (1/2, 1/2)
        super(HawkDove, self).__init__(payoff_matrices=payoff_matrix, player_frequencies=player_dist,      equilibrium_tolerance=equilibrium_tolerance)

    @classmethod
    def classify(cls, params, state, tolerance):
        p = params
        threshold = 1-tolerance
    
        if state[0][0] >= threshold:
            return 0
        elif state[1][1] >= threshold:
            return 1
        elif state[1][0] >= threshold:
            return 2
        elif state[0][1] >= threshold:
           return 3
       else:
           return super(HawkDove, cls).classify(params, state, tolerance)

• Determine which game you wish to model and subclass the game class.
• Create the defaults values and parameters.
• Prepare the payoff matrices for each player they should be n-dimensional matrices which vary player 1's strategy at the first level, then player 2's strategy at the second level etc...
• Call the super class method with the matrices, player distributions, and equilibria.
• Override the classification of equilibria class method, to determine when equilibria are achieved. The first index of the variable 'state' determines the player, and the second index determine their strategy.