def nearby_profs(self, prof, num_devs):
"""Returns profiles reachable by at most num_devs deviations"""
# XXX this is the bottleneck for gpgame.neighbor_EVs. It seems like
# there should be some clever way to speed it up.
assert num_devs >= 0
dev_players = utils.acomb(self.num_roles, num_devs, True)
mask = np.all(dev_players <= self.num_players, 1)
dev_players = dev_players[mask]
supp = prof > 0
sub = subgame.subgame(rsgame.BaseGame(self), supp)
profs = [prof[None]]
for players in dev_players:
to_dev_profs = rsgame.BaseGame(
players, self.num_strategies).all_profiles()
from_dev_profs = subgame.translate(
rsgame.BaseGame(players, sub.num_strategies).all_profiles(),
supp)
before_devs = prof - from_dev_profs
before_devs = before_devs[np.all(before_devs >= 0, 1)]
before_devs = utils.unique_axis(before_devs)
nearby = before_devs[:, None] + to_dev_profs
nearby.shape = (-1, self.num_role_strats)
profs.append(utils.unique_axis(nearby))
profs = np.concatenate(profs)
return utils.unique_axis(profs)
def test_acomb():
actual = utils.acomb(5, 0)
assert actual.shape == (1, 5)
assert not actual.any()
actual = utils.acomb(5, 5)
assert actual.shape == (1, 5)
assert actual.all()
actual = utils.acomb(6, 4)
expected = np.zeros_like(actual)
for i, inds in enumerate(itertools.combinations(range(6), 4)):
expected[i, inds] = True
assert array_set_equals(actual, expected)
actual = utils.acomb(6, 4, True)
expected = np.zeros_like(actual)
for i, inds in enumerate(map(list, itertools.combinations_with_replacement(
range(6), 4))):
np.add.at(expected[i], inds, 1)
assert array_set_equals(actual, expected)
def test_acomb():
"""Test acomb"""
actual = utils.acomb(5, 0)
assert actual.shape == (1, 5)
assert not actual.any()
actual = utils.acomb(5, 5)
assert actual.shape == (1, 5)
assert actual.all()
actual = utils.acomb(6, 4)
expected = np.zeros_like(actual)
for i, inds in enumerate(itertools.combinations(range(6), 4)):
expected[i, inds] = True
assert array_set_equals(actual, expected)
actual = utils.acomb(6, 4, True)
expected = np.zeros_like(actual)
for i, inds in enumerate(
map(list, itertools.combinations_with_replacement(range(6), 4))):
np.add.at(expected[i], inds, 1)
assert array_set_equals(actual, expected)
def grid_mixtures(self, num_points):
"""Returns all of the mixtures in a grid with n points
Arguments
---------
num_points : int > 1
The number of points to have along one dimensions
"""
assert num_points > 1, "Must have at least two points on a dimensions"
role_mixtures = [utils.acomb(num_strats, num_points - 1, True) /
(num_points - 1)
for num_strats in self.num_strategies]
return utils.acartesian2(*role_mixtures)
def test_acomb():
actual = utils.acomb(5, 0)
assert actual.shape == (1, 5)
assert not actual.any()
actual = utils.acomb(5, 5)
assert actual.shape == (1, 5)
assert actual.all()
actual = utils.acomb(6, 4)
expected = np.zeros_like(actual)
for i, inds in enumerate(itertools.combinations(range(6), 4)):
expected[i, inds] = True
assert np.setxor1d(utils.axis_to_elem(actual),
utils.axis_to_elem(expected)).size == 0
actual = utils.acomb(6, 4, True)
expected = np.zeros_like(actual)
for i, inds in enumerate(map(list, itertools.combinations_with_replacement(
range(6), 4))):
np.add.at(expected[i], inds, 1)
assert np.setxor1d(utils.axis_to_elem(actual),
utils.axis_to_elem(expected)).size == 0
def congestion(num_players, num_facilities, num_required, degree=2,
coef_dist=lambda d: -np.random.exponential(10. ** (1 - d))):
"""Generate a congestion game
Parameters
----------
num_players : int
num_facilities : int
num_required : int
degree : int, optional
Degree of payoff polynomials
coef_dist : f(int) -> float, optional
Numpy compatible function for generating random coefficients
conditioned on degree. Leading degree must be negative to generate a
true congestion game.
"""
function_inputs = utils.acomb(num_facilities, num_required)
table_dist = random_poly_dist(coef_dist, np.insert(np.zeros(degree), 2, 1))
functions = table_dist((num_facilities, num_players + 1))
return aggfn.aggfn(num_players, function_inputs.shape[0],
function_inputs.T, function_inputs, functions)
def congestion(num_players, num_facilities, num_required, *, degree=2):
"""Generate a congestion game
A congestion game is a symmetric game, where there are a given number of
facilities, and each player must choose to use some amount of them. The
payoff for each facility decreases as more players use it, and a players
utility is the sum of the utilities for every facility.
In this formulation, facility payoffs are random polynomials of the number
of people using said facility.
Parameters
----------
num_players : int > 1
The number of players.
num_facilities : int > 1
The number of facilities.
num_required : 0 < int < num_facilities
The number of required facilities.
degree : int > 0, optional
Degree of payoff polynomials.
"""
utils.check(num_players > 1, 'must have more than one player')
utils.check(num_facilities > 1, 'must have more than one facility')
utils.check(
0 < num_required < num_facilities,
'must require more than zero but less than num_facilities')
utils.check(degree > 0, 'degree must be greater than zero')
function_inputs = utils.acomb(num_facilities, num_required)
functions = -_random_monotone_polynomial(num_facilities, num_players,
degree)
facs = tuple(utils.prefix_strings('', num_facilities))
strats = tuple('_'.join(facs[i] for i, m in enumerate(mask) if m)
for mask in function_inputs)
return aggfn.aggfn_names(
['all'], num_players, [strats], function_inputs.T, function_inputs,
functions)
def all_profiles(self):
"""Return all profiles"""
role_arrays = [utils.acomb(n_strats, players, True)
for n_strats, players
in zip(self.num_strategies, self.num_players)]
return utils.acartesian2(*role_arrays)