def minimize_pairwise_connectivity(G, S0):
tmp = S0.copy()
nodes = tmp.copy()
node = next(iter(tmp))
tmp.discard(node)
subgraph = nx.subgraph_view(G, filter_node=lambda n: n not in tmp)
connectivity = connectivity_metrics.pairwise_connectivity(subgraph)
vertices = [node]
minimum = connectivity
while tmp:
node = next(iter(tmp))
tmp.discard(node)
nodes.discard(node)
subgraph = nx.subgraph_view(G, filter_node=lambda n: n not in nodes)
connectivity = connectivity_metrics.pairwise_connectivity(subgraph)
if connectivity < minimum:
vertices = [node]
minimum = connectivity
elif connectivity == minimum:
vertices.append(node)
nodes.add(node)
return vertices
def maximize_disconnected_pairs(G, k):
tmp = set(G)
nodes = tmp.copy()
node = next(iter(tmp))
tmp.discard(node)
subgraph = nx.subgraph_view(G, filter_node=lambda n: n in tmp)
connectivity = comb(
len(tmp), k) - connectivity_metrics.pairwise_connectivity(subgraph)
vertices = [node]
maximum = connectivity
while tmp:
node = next(iter(tmp))
tmp.discard(node)
nodes.discard(node)
subgraph = nx.subgraph_view(G, filter_node=lambda n: n in nodes)
connectivity = comb(
len(tmp), k) - connectivity_metrics.pairwise_connectivity(subgraph)
if connectivity > maximum:
vertices = [node]
maximum = connectivity
elif connectivity == maximum:
vertices.append(node)
nodes.add(node)
return vertices
def get_pairwise_connectivity(exclude=None):
if exclude is None:
exclude = {}
S = set(exclude)
subgraph = nx.subgraph_view(G, filter_node=lambda n: n not in S)
return connectivity_metric.pairwise_connectivity(subgraph)
def get_critical_nodes_ga(G, k):
S, _ = cnp.genetic_algorithm(G, k)
subgraph = nx.subgraph_view(G, filter_node=lambda n: n not in S)
fitness = connectivity_metrics.pairwise_connectivity(subgraph)
print(fitness)
return fitness
def get_critical_nodes_greedy(G, k):
S = cnp.greedy_cnp(G, k)
subgraph = nx.subgraph_view(G, filter_node=lambda n: n not in S)
fitness = connectivity_metrics.pairwise_connectivity(subgraph)
print(fitness)
return fitness
def _update(G, best_S, P, alpha):
subgraph = nx.subgraph_view(G, filter_node=lambda n: n not in best_S)
metric = connectivity_metric.pairwise_connectivity(subgraph)
avg = 0
for S in P:
avg += len(S.intersection(best_S))
avg /= len(P)
try:
gamma = (alpha * metric) / avg
except:
gamma = 1
return gamma
def fitness_function(S):
subgraph = nx.subgraph_view(G, filter_node=lambda n: n not in S)
metric = connectivity_metric.pairwise_connectivity(subgraph)
commonalities = S.intersection(best_S)
return metric + gamma * len(commonalities)