PayoffMatrix

tomsup.payoffmatrix

This scripts contains the PayoffMatrix

class tomsup.payoffmatrix.PayoffMatrix(name: str, predefined: Optional[array] = None)

Bases: object

A class of 2 by 2 payoff matrices.

Currently include the following games: The staghunt game: ‘staghunt’, The matching pennies game (coop and competive): ‘penny_competive’, ‘penny_cooperative’, The party dilemma: ‘party’, The Battle of the sexes: ‘sexes’, The chicken game: ‘chicken’, The deadlock: ‘deadlock’, The prisoners dilemma: ‘prisoners_dilemma’.

For custom payoff matrix supply a 2x2x2 numpy array to the predefined argument.

Examples

>>> import tomsup as ts
>>> staghunt = ts.PayoffMatrix(name="staghunt")
>>> staghunt.payoff(action_agent0=1, action_agent1=1, agent=0)
5
>>> staghunt.payoff(action_agent0=1, action_agent1=0, agent=0)
0
>>> staghunt.payoff(action_agent0=0, action_agent1=1, agent=0)
3
>>> chicken = ts.PayoffMatrix(name="chicken")
>>> chicken.payoff(0, 1, 0)
-1
>>> dead = ts.PayoffMatrix(name="deadlock")
>>> dead.payoff(1, 0, 1)
0
>>> sexes = ts.PayoffMatrix(name="sexes")
>>> sexes.payoff(1, 1, 0)
5
>>> custom = ts.PayoffMatrix(name="custom", np.array(([(10, 0), (0, 5)],
                                                [(5, 0), (0, 10)])))

get_matrix() → array

Returns:

The payoff matrix

Return type:

np.array

payoff(choice_agent0: int, choice_agent1: int, agent: int = 0) → float

Parameters:

• choice_agent0 (int) – choice of agent 0

• choice_agent1 (int) – choice of agent 1

• agent (int, optional) – The perspective agent which should get the payoff, either 0 or 1. Defaults to 0.

Returns:

The payoff of the agent

Return type:

float