import numpy as np
import matplotlib.pyplot as plt
from tabulate import tabulate
from elsim.methods import (fptp, borda, utility_winner, approval)
from elsim.elections import random_utilities
from elsim.strategies import honest_rankings, vote_for_k
from weber_1977_expressions import eff_standard, eff_vote_for_half, eff_borda
n = 2_000
n_voters = 1_000
n_cands_list = (2, 3, 4, 5, 6, 10, 255)
ranked_methods = {'Standard': fptp, 'Borda': borda}
rated_methods = {'Vote-for-half': lambda utilities, tiebreaker:
approval(vote_for_k(utilities, 'half'), tiebreaker)}
count = {key: Counter() for key in (ranked_methods.keys() |
rated_methods.keys() | {'UW'})}
start_time = time.monotonic()
for iteration in range(n):
for n_cands in n_cands_list:
utilities = random_utilities(n_voters, n_cands)
# Find the social utility winner and accumulate utilities
UW = utility_winner(utilities)
count['UW'][n_cands] += utilities.sum(axis=0)[UW]
for name, method in rated_methods.items():
from joblib import Parallel, delayed
from elsim.methods import (fptp, utility_winner, approval)
from elsim.elections import random_utilities
from elsim.strategies import honest_rankings, vote_for_k
from weber_1977_expressions import (eff_standard, eff_vote_for_k,
eff_vote_for_half)
n = 10_000
n_voters = 1_000
n_cands_list = np.arange(2, 11)
ranked_methods = {'Standard': fptp}
rated_methods = {
'Vote-for-1':
lambda utilities, tiebreaker: approval(vote_for_k(utilities, 1), tiebreaker
),
'Vote-for-2':
lambda utilities, tiebreaker: approval(vote_for_k(utilities, 2), tiebreaker
),
'Vote-for-3':
lambda utilities, tiebreaker: approval(vote_for_k(utilities, 3), tiebreaker
),
'Vote-for-4':
lambda utilities, tiebreaker: approval(vote_for_k(utilities, 4), tiebreaker
),
'Vote-for-half':
lambda utilities, tiebreaker: approval(vote_for_k(utilities, 'half'),
tiebreaker),
'Vote-for-(n-1)':
lambda utilities, tiebreaker: approval(vote_for_k(utilities, -1),
tiebreaker),
D = 2
ranked_methods = {
'Plurality': fptp,
'Runoff': runoff,
'Hare': irv,
'Borda': borda,
'Coombs': coombs,
'Black': black
}
rated_methods = {
'SU max':
utility_winner,
'Approval':
lambda utilities, tiebreaker: approval(approval_optimal(utilities),
tiebreaker)
}
# Plot Merrill's results as dotted lines for comparison (traced from plots)
merrill_fig_2c = {
'Black': {
2: 100.0,
3: 100.0,
4: 100.0,
5: 100.0,
7: 100.0
},
'Coombs': {
2: 100.0,
3: 99.1,
4: 98.1,
from tabulate import tabulate
from elsim.methods import (fptp, runoff, irv, approval, borda, coombs,
black, utility_winner, condorcet)
from elsim.elections import random_utilities
from elsim.strategies import honest_rankings, approval_optimal
n = 10_000
n_voters = 25
n_cands_list = (2, 3, 4, 5, 7, 10)
ranked_methods = {'Plurality': fptp, 'Runoff': runoff, 'Hare': irv,
'Borda': borda, 'Coombs': coombs, 'Black': black}
rated_methods = {'SU max': utility_winner,
'Approval': lambda utilities, tiebreaker:
approval(approval_optimal(utilities), tiebreaker)}
count = {key: Counter() for key in (ranked_methods.keys() |
rated_methods.keys() | {'CW'})}
start_time = time.monotonic()
for iteration in range(n):
for n_cands in n_cands_list:
utilities = random_utilities(n_voters, n_cands)
"""
"Simulated utilities were normalized by range, that is, each voter's
set of utilities were linearly expanded so that the highest and lowest
utilities for each voter were 1 and 0, respectively."