def iteration(g, v):
"""
Performs a single iteration of the simulation.
:param g: Graph `G = (M, C)`.
:param v: Set of votes `V` (excluding 0).
"""
# Activation ratio of a channel `x`.
def a(x):
return g.edges[x]['a']
c = {c for c in g.edges if trial(a(c))} # Active channels.
i = {m: neighbours(m, c) for m in g.nodes} # Active neighbours.
w = {(m, n): valuation_density(g, m, n) for m in g.nodes for n in i[m]}
for m, data in g.nodes(data=True):
if len(i[m]) == 0:
# Skip members that did not have active
# neighbours to communicate with.
continue
# Optimisation to prevent multiple evaluations.
cached = functools.lru_cache(maxsize=len(v))
data['w'] = cached(average_density(m, i, w))
data['d'] = decision(v, data['w'], data['ro'])
def nodes(g, v, ro):
"""
Initialises attributes of the nodes of the graph `g`:
- 'w' for preference density function;
- 'd' for initial decision of a member;
- 'ro' for impulsiveness indicator.
:param g: Graph.
:param v: Set of votes `V` (excluding 0).
:param ro: Function that returns impulsiveness indicator.
Examples:
nodes(g, v, ro=lambda: 25)
nodes(g, v, ro=lambda: uniform(20, 30))
"""
for _, data in g.nodes(data=True):
w = {v: int(uniform() * 10000) for v in v}
w = {v: w[v] / sum(w.values()) for v in v}
w = function(w)
data['w'] = w # Preference density function.
data['ro'] = ro() # Impulsiveness indicator.
data['d'] = decision(v, data['w'], data['ro']) # Initial decision.
assert sum(w(v) for v in v) == 1
assert data['ro'] > 0
assert data['d'] == 0 or data['d'] in v
def leader(g, v, m, j, a=lambda: 1.0):
"""
Makes a leader out of the member.
:param g: Graph.
:param v: Set of votes `V` (excluding 0).
:param m: The node that will become a leader.
:param j: Vote of the leader.
:param a: Channel activation function.
"""
w = {v: uniform(0.001, 0.002) for v in v}
w[j] = 1 - sum(w.values()) + w[j]
w = function(w)
data = g.nodes[m]
data['w'] = w # Preference density function.
data['d'] = decision(v, data['w'], data['rho']) # Initial decision.
for n in g.neighbors(m):
r = uniform(0, 0.5)
data = g.edges[m, n]
data['a'] = a() # Activation ratio.
data['d'] = { # Dialogue matrix.
(0, 0): r,
(0, 1): 1 - r,
(1, 0): 0,
(1, 1): 0,
}