A computational system for analyzing economic mechanisms
To solve games with additional solvers, download the AGG solvers at the AGG website
Also see the PosEc website for links to publications.
Questions, bug reports and feature suggestions should be directed to Neil Newman - newmanne at cs dot ubc dot ca
PosEc is the product of the ideas and hard work of David Thompson, Neil Newman, and Kevin Leyton-Brown.
Here is the code needed to define a plurality voting game.
import itertools
from posec import *
#
# Setting
#
# 10 agents
n = 10
# 3 outcomes
O = ("A", "B", "C")
# Every preference ordering is possible
Theta = [list(itertools.permutations(O))]
# Every agent is equally likely to have every preference ordering
P = [UniformDistribution(Theta)] * n
# Utility function - 0 points if least preferred candidate wins, 1 if next most preferred etc.
def u(i, theta, o, a_i):
return theta[i].index(o)
# Create the setting object
s = BayesianSetting(n, O, Theta, P, u)
#
# Mechanism
#
# Action function: the possible actions for each player (here to vote for any candidate)
def A(setting, i, theta_i):
return setting.O
# The choice function - given all of the agents' actions, return a distribution over outcomes
def M(setting, a_N):
scores = {o: a_N.count(o) for o in setting.O}
maxScore = max(scores.values())
winners = [o for o in scores.keys() if scores[o] == maxScore]
return UniformDistribution(winners)
# Create the mechanism object
m = Mechanism(A, M)
# Run our structure inference algorithms to make a BAGG
agg = makeAGG(s, m, symmetry=True)
# Save it to a file for use with any tool that supports BAGGs!
agg.saveToFile('plurality.bagg')Be sure to check out our examples section that defines mechanisms and settings for position auctions and voting games. You are encouraged to contribute further mechanisms and settings!
- First official release of PosEc.