def fastTargetedOrder(graph):
''' input : graph -- a graph
output: a list, of nodes in decreasing order of their degrees
'''
new_graph = provided.copy_graph(graph)
number_nodes = len(new_graph.keys())
degree_sets = []
for dummy in range(number_nodes):
degree_sets.append(set([]))
for node in new_graph.keys():
#print node, new_graph[node]
degree = len(new_graph[node])
#print degree
degree_sets[degree].update([node])
#print '*****'
#for item in degree_sets:
#print item
order = []
for degree in range(number_nodes -1, -1, -1):
#print 'Degree: ', degree
while degree_sets[degree]:
#print new_graph
node = degree_sets[degree].pop()
#print type(new_graph[node])
#print new_graph#[node]
for neighbor in new_graph[node]:
neighbor_degree = len(new_graph[neighbor])
degree_sets[neighbor_degree].remove(neighbor)
degree_sets[neighbor_degree - 1].update([neighbor])
order.append(node)
provided.delete_node(new_graph,node)
return order
def fast_targeted_order(ugraph):
"""
Compute a targeted attack order consisting of nodes of maximal degree.
"""
n = len(ugraph)
new_graph = app2.copy_graph(ugraph)
degree_sets = {}
for node, neighbors in ugraph.iteritems():
degree_sets.setdefault(len(neighbors), set()).add(node)
order = []
i = 0
for k in range(n-1, 0, -1):
if degree_sets.has_key(k):
while len(degree_sets[k]) != 0:
max_degree_node = degree_sets[k].pop()
for neighbor in new_graph[max_degree_node]:
d = len(new_graph[neighbor])
degree_sets[d].remove(neighbor)
degree_sets.setdefault(d-1, set()).add(neighbor)
order.append(max_degree_node)
prj2.delete_node(new_graph, max_degree_node)
i += 1
if i < n:
order.extend(new_graph.keys())
return order
def fast_targeted_order(ugraph):
"""
Compute a targeted attack order consisting of nodes of maximal degree.
"""
n = len(ugraph)
new_graph = app2.copy_graph(ugraph)
degree_sets = {}
for node, neighbors in ugraph.iteritems():
degree_sets.setdefault(len(neighbors), set()).add(node)
order = []
i = 0
for k in range(n - 1, 0, -1):
if degree_sets.has_key(k):
while len(degree_sets[k]) != 0:
max_degree_node = degree_sets[k].pop()
for neighbor in new_graph[max_degree_node]:
d = len(new_graph[neighbor])
degree_sets[d].remove(neighbor)
degree_sets.setdefault(d - 1, set()).add(neighbor)
order.append(max_degree_node)
prj2.delete_node(new_graph, max_degree_node)
i += 1
if i < n:
order.extend(new_graph.keys())
return order
def fast_targeted_order(ugraph): ugraph = provided.copy_graph(ugraph) n = len(ugraph) degree_sets = dict() for k in range(n): degree_sets[k] = set([]) for node in range(n): node_degree = len(ugraph[node]) degree_sets[node_degree].add(node) L = list() for k in range(n - 1, -1, -1): while len(degree_sets[k]) > 0: u = degree_sets[k].pop() for neighbor in ugraph[u]: degree = len(ugraph[neighbor]) degree_sets[degree].remove(neighbor) degree_sets[degree-1].add(neighbor) L.append(u) provided.delete_node(ugraph, u) return L
def compute_resilience(ugraph, attack_order):
'''Compute graph resilience (size of the largest connected component)
after removing each of provided nodes.'''
ugraph = alg_application2_provided.copy_graph(ugraph)
resilience = [largest_cc_size(ugraph)]
for node in attack_order:
remove_node(ugraph, node)
resilience.append(largest_cc_size(ugraph))
return resilience
def compute_resilience(ugraph, attack_order):
'''Compute graph resilience (size of the largest connected component)
after removing each of provided nodes.'''
ugraph = alg_application2_provided.copy_graph(ugraph)
resilience = [largest_cc_size(ugraph)]
for node in attack_order:
remove_node(ugraph, node)
resilience.append(largest_cc_size(ugraph))
return resilience
def attack(graph, attack_type):
new_graph = provided.copy_graph(graph)
if attack_type == 'normal':
order = provided.targeted_order(new_graph)
elif attack_type == 'fast':
order = fastTargetedOrder(new_graph)
resilience = pa2.compute_resilience(new_graph, order)
return resilience
#t0 = time.time()
#normal_resilience = attack(upa, 'normal')
#normal_time = time.time() - t0
#
#t0 = time.time()
#fast_resilience = attack(upa, 'fast')
#fast_time = time.time() - t0
##
#print 'Normal attack time: ', normal_time
#print 'Fast attack time: ', fast_time
#print fastTargetedOrder(create_UPA(25,3))
def fast_targeted_order(ugraph):
ugraph = alg_application2_provided.copy_graph(ugraph)
N = len(ugraph)
degree_sets = [set()] * N
for node, neighbors in ugraph.iteritems():
degree = len(neighbors)
degree_sets[degree].add(node)
order = []
for k in range(N - 1, -1, -1):
while degree_sets[k]:
u = degree_sets[k].pop()
for neighbor in ugraph[u]:
d = len(ugraph[neighbor])
degree_sets[d].remove(neighbor)
degree_sets[d - 1].add(neighbor)
order.append(u)
project2.remove_node(ugraph, u)
return order
def fast_targeted_order(ugraph):
ugraph = alg_application2_provided.copy_graph(ugraph)
N = len(ugraph)
degree_sets = [set()] * N
for node, neighbors in ugraph.iteritems():
degree = len(neighbors)
degree_sets[degree].add(node)
order = []
for k in range(N - 1, -1, -1):
while degree_sets[k]:
u = degree_sets[k].pop()
for neighbor in ugraph[u]:
d = len(ugraph[neighbor])
degree_sets[d].remove(neighbor)
degree_sets[d - 1].add(neighbor)
order.append(u)
module2.remove_node(ugraph, u)
return order
def fast_targeted_order(ugraph):
new_graph = alg_application2_provided.copy_graph(ugraph)
DegreeSets = degreesets(new_graph)
#print DegreeSets
#print new_graph
L = list()
i = 0
for k in range(len(new_graph)-1,-1,-1):
while DegreeSets[k] != set():
u = DegreeSets[k].pop()
#print u
#print new_graph[u]
for neighbor in new_graph[u]:
d = len(new_graph[neighbor])
DegreeSets[d].remove(neighbor)
DegreeSets[d-1].add(neighbor)
#print "DegreeSets", DegreeSets
L.insert(i,u)
i = i + 1
#if new_graph.get(u):
alg_application2_provided.delete_node(new_graph,u)
return L
#graph of ER
er_graph = er_algorithm(len(cn_graph.keys()), 0.002)
#print er_graph
print edges_in_undirected_graph(er_graph)
#graph of UPA
upa_graph = get_upa_graph(len(cn_graph.keys()), 2)
print edges_in_undirected_graph(upa_graph)
#random attack
cn_attack_order = random_order(cn_graph)
er_attack_order = random_order(er_graph)
upa_attack_order = random_order(upa_graph)
#resilience for different graphs
cn_resilience = pj2.compute_resilience(aa2p.copy_graph(cn_graph), cn_attack_order)
er_resilience = pj2.compute_resilience(aa2p.copy_graph(er_graph), er_attack_order)
upa_resilience = pj2.compute_resilience(aa2p.copy_graph(upa_graph), upa_attack_order)
#making comparison plots
fig, ax = plt.subplots()
print len(cn_resilience)
x_vals = [idx for idx in range(len(cn_graph.keys()) + 1)]
print len(x_vals)
ax.plot(x_vals, cn_resilience, '-b', label = 'computer network')
ax.plot(x_vals, er_resilience, '-r', label = 'ER graph (p = 0.002)')
ax.plot(x_vals, upa_resilience, '-g', label = 'UPA graph (m = 2)')
ax.legend(loc = 'upper right')
ax.set_title('Resiliences of different models ')
import algo_app2_1 as app2
import alg_application2_provided as provided
import algo_pa2 as pa2
import pickle
create_upas = []
normal_time = []
fast_time = []
for n in range(10, 1000, 10):
print n
t0 = time.time()
upa = app2.create_UPA(n, 5)
create_upas.append(time.time() - t0)
#print upa
use_graph = provided.copy_graph(upa)
t0 = time.time()
order = provided.targeted_order(use_graph)
#normal_resilience = pa2.compute_resilience(use_graph, order)
#normal_resilience = app2.attack(upa, 'normal')
normal_time.append(time.time() - t0)
use_graph = provided.copy_graph(upa)
t0 = time.time()
order = app2.fastTargetedOrder(use_graph)
#fast_resilience = pa2.compute_resilience(use_graph, order)
#fast_resilience = app2.attack(upa, 'fast')
fast_time.append(time.time() - t0)
