def averageNodeData():
sizeOfCover = range(0,200)
averageRemovedEdges = range(0,200)
G = nx.erdos_renyi_graph(100,0.05)
temp_data = range(0,200)
for j in range(0,200):
process = Popen(["../build/programming3", utils.dmax(G),"r"], stdout=PIPE)
(output, err) = process.communicate()
print(output)
exit_code = process.wait()
data = re.findall(r"[-+]?\d*\.\d+|\d+",output)
data = [float(s) for s in data]
print(data)
sizeOfCover[j] = data[0]
averageRemovedEdges[j] = data[1]
plt.plot(averageRemovedEdges, sizeOfCover, 'ro')
plt.ylabel('Size Vertex Cover')
plt.xlabel('Average #Edge Deletions per Iteration')
plt.savefig('average_node_deletions.png', dpi=100)
def histogram():
for i in range(1,10):
G=nx.erdos_renyi_graph(100,0.01*i)
temp_data = range(0,500)
for j in range(0,500):
process = Popen(["../build/programming3", utils.dmax(G),"r"], stdout=PIPE)
(output, err) = process.communicate()
exit_code = process.wait()
data = re.findall(r"[-+]?\d*\.\d+|\d+",output)
data = [float(s) for s in data]
temp_data[j]=data[0]
plt.clf()
pl.clf()
(n,bins,patches)=pl.hist(temp_data,5,normed=1,histtype='bar',color='Burlywood',label='Histogram '+str(0.2*i))
y = plt.mlab.normpdf(bins,np.mean(temp_data),np.std(temp_data))
pl.plot(bins,y,'k--',linewidth=1.5,label='Fitted Gaussian')
plt.legend()
plt.xlabel('Size Vertex Cover')
plt.ylabel('Probability')
plt.savefig('histogram' + str(0.2*i) + '.png', dpi=100)
def generate_data():
apxg_vc = range(0,10)
apxi_vc = range(0,10)
apxir_vc = range(0,10)
apxr_vc = range(0,10)
apxr_vc_mean = range(0,10)
apxr_vc_errorl = range(0,10)
apxr_vc_errorh = range(0,10)
xaxis = range(0,10)
ind = range(0,10)
for i in range(0,10):
print(str(i+1)+" ")
#G=nx.erdos_renyi_graph(200,0.01*(i+1))
#xaxis[i] = 0.01*(i+1)
G = nx.random_regular_graph((i+1)*10, 101)
xaxis[i] = (i+1)*10
#G = nx.powerlaw_cluster_graph(50, i+2, 0.0)
temp_data = range(0,100)
for j in range(0,100):
process = Popen(["../build/programming3", utils.dmax(G),"r"], stdout=PIPE)
(output, err) = process.communicate()
exit_code = process.wait()
data = re.findall(r"[-+]?\d*\.\d+|\d+",output)
data = [float(s) for s in data]
temp_data[j]=data[0]
#Random
apxr_vc_mean[i] = np.mean(temp_data)
apxr_vc_errorl[i] = np.min(temp_data)-np.mean(temp_data)
apxr_vc_errorh[i] = np.mean(temp_data)-np.max(temp_data)
apxr_vc[i] = temp_data
#Greedy
process = Popen(["../build/programming3", utils.dmax(G),"g"], stdout=PIPE)
(output, err) = process.communicate()
exit_code = process.wait()
apxg_vc[i] = int(output)
#Greedy Improved
process = Popen(["../build/programming3", utils.dmax(G),"i"], stdout=PIPE)
(output, err) = process.communicate()
exit_code = process.wait()
apxi_vc[i] = int(output)
#Greedy Improved Reversed
process = Popen(["../build/programming3", utils.dmax(G),"!"], stdout=PIPE)
(output, err) = process.communicate()
exit_code = process.wait()
apxir_vc[i] = int(output)
plt.errorbar(ind,apxr_vc_mean,[apxr_vc_errorl,apxr_vc_errorh],fmt='-bo')
plt.ylabel('Size Vertex Cover')
plt.xlabel('% Graph Connectivity')
#Random and Greedy
b1 = plt.plot(xaxis,apxr_vc_mean,'-kx',label='Random Heuristic')
plt.boxplot(apxr_vc, sym='k+', positions=xaxis,notch=1)
b3 = plt.plot(xaxis,apxg_vc,'-.ro',label='Greedy Heuristic')
plt.legend(loc='lower right')
plt.savefig('RandomGreedy.png', dpi=100)
#With pruning
b4 = plt.plot(xaxis,apxi_vc,'-g^',label='Greedy Heuristic & Min Greedy Cover Pruning')
b5 = plt.plot(xaxis,apxir_vc,'--bs',label='Greedy Heuristic & Max Greedy Cover Pruning')
plt.legend(loc='lower right')
plt.savefig('Pruning.png', dpi=100)