-
Notifications
You must be signed in to change notification settings - Fork 6
/
figure_9.py
258 lines (221 loc) · 10 KB
/
figure_9.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
import os
import argparse
import matplotlib.pyplot as plt
import pickle
import numpy as np
def frange(x, y, jump):
while x < y:
yield x
x += jump
def read_defo_optim_time(standard_out_file):
optim_time = 0
with open(standard_out_file) as fd:
while (True):
line = fd.readline()
if line.startswith("optimization time"):
camps = line.split(":")
optim_time = float(camps[-1].split('\n')[0])
break
return float(optim_time)/1000
def read_max_load_link(standard_out_file):
pre_optim_max_load_link, post_optim_max_load_link = 0, 0
with open(standard_out_file) as fd:
while (True):
line = fd.readline()
if line.startswith("pre-optimization"):
camps = line.split(" ")
pre_optim_max_load_link = float(camps[-1].split('\n')[0])
elif line.startswith("post-optimization"):
camps = line.split(" ")
post_optim_max_load_link = float(camps[-1].split('\n')[0])
break
return (pre_optim_max_load_link, post_optim_max_load_link)
if __name__ == "__main__":
# This script is to obtain the figure 10 from Conext 2021 paper.
# Before executing this file we must execute the eval_on_zoo_topologies.py file to evaluate the DRL model and store the results
# python figure_9.py -d SP_3top_15_B_NEW -p ../rwds-results-1-link_capacity-unif-05-1-zoo
parser = argparse.ArgumentParser(description='Parse files and create plots')
# The flag 'p' points to the folder where the .pckl files are found
parser.add_argument('-p', help='data folder', type=str, required=True, nargs='+')
parser.add_argument('-d', help='differentiation string for the model', type=str, required=True, nargs='+')
args = parser.parse_args()
differentiation_str = args.d[0]
directory = args.p[0]+'/'+differentiation_str
uti_sim_anneal = []
uti_SAP = []
uti_DRL_SP = []
uti_MLP = []
uti_ALL = []
uti_DRL_HILL = []
uti_HILL = []
uti_DEFO = []
cost_DEFO = []
cost_SAP = []
cost_DRL_SP = []
cost_DRL_HILL = []
cost_HILL = []
filename_list = []
X_axis = []
if not os.path.exists("./Images"):
os.makedirs("./Images")
path_to_dir = "./Images/EVALUATION/"+differentiation_str+'/'
if not os.path.exists(path_to_dir):
os.makedirs(path_to_dir)
aux_path_to_DEFO_results = "../Enero_datasets/DEFOResults/results-1-link_capacity-unif-05-1-zoo/"
axis_counter = 0
# Iterate over all topologies and evaluate our DRL agent on all TMs
for filename in os.listdir(directory):
dir_to_topology_rewards = directory+"/"+filename
aux_sim_anneal = []
aux_hill_climb = []
aux_SAP = []
aux_MLP = []
aux_DRL_SP = []
aux_DRL_HILL = []
aux_DEFO = []
aux_cost_SAP = []
aux_cost_DEFO = []
aux_cost_DRL_SP = []
aux_cost_DRL_HILL = []
aux_cost_hill_climb = []
for subdir, dirs, files in os.walk(dir_to_topology_rewards):
for file in files:
if file.endswith((".pckl")):
top_name = file.split('.')[0]
tm_id = file.split('.')[1]
path_to_DEFO_file_results = aux_path_to_DEFO_results+top_name+"/standard_out_"+top_name+"_"+tm_id
defo_time = read_defo_optim_time(path_to_DEFO_file_results)
results = []
path_to_pckl_rewards = dir_to_topology_rewards + '/'
with open(path_to_pckl_rewards+file, 'rb') as f:
results = pickle.load(f)
aux_sim_anneal.append(results[4])
aux_hill_climb.append(results[7])
aux_SAP.append(results[8])
aux_MLP.append(results[5])
aux_DRL_SP.append(results[9])
aux_DEFO.append(results[1])
aux_DRL_HILL.append(results[3])
aux_cost_SAP.append(results[13])
aux_cost_DEFO.append(defo_time)
aux_cost_DRL_SP.append(results[14])
aux_cost_DRL_HILL.append(results[16])
aux_cost_hill_climb.append(results[15])
sim_anneal_mean = np.mean(aux_hill_climb) # np.mean(aux_sim_anneal)
uti_SAP.append((sim_anneal_mean-np.mean(aux_SAP))/sim_anneal_mean)
uti_MLP.append((sim_anneal_mean-np.mean(aux_MLP))/sim_anneal_mean)
uti_DRL_SP.append((sim_anneal_mean-np.mean(aux_DRL_SP))/sim_anneal_mean)
uti_DRL_HILL.append((sim_anneal_mean-np.mean(aux_DRL_HILL))/sim_anneal_mean)
uti_HILL.append((sim_anneal_mean-np.mean(aux_hill_climb))/sim_anneal_mean)
uti_DEFO.append((sim_anneal_mean-np.mean(aux_DEFO))/sim_anneal_mean)
cost_DEFO.append(np.mean(aux_cost_DEFO))
cost_SAP.append(np.mean(aux_cost_SAP))
cost_DRL_SP.append(np.mean(aux_cost_DRL_SP))
cost_DRL_HILL.append(np.mean(aux_cost_DRL_HILL))
cost_HILL.append(np.mean(aux_cost_hill_climb))
X_axis.append(axis_counter)
filename_list.append(filename)
axis_counter += 1
# We do the following to order the scores
for i in range(len(X_axis)):
uti_ALL.append((filename_list[i], uti_SAP[i], uti_DRL_SP[i], uti_DRL_HILL[i]-uti_HILL[i], uti_DRL_HILL[i], cost_SAP[i], cost_DRL_SP[i], cost_DRL_HILL[i], uti_HILL[i], cost_HILL[i], uti_DEFO[i], cost_DEFO[i], 1, 0))
new_uti_ALL = sorted(uti_ALL, key=lambda tup: tup[3], reverse=False)
print(new_uti_ALL)
print(len(new_uti_ALL))
dict_tops = dict()
dict_tops["BtAsiaPac"] = 1
dict_tops["Goodnet"] = 1
dict_tops["Garr199905"] = 1
for i in range(len(X_axis)):
if new_uti_ALL[i][0] in dict_tops:
print(i, new_uti_ALL[i][0])
uti_SAP = []
uti_DRL_SP = []
uti_DRL_HILL = []
uti_HILL = []
uti_DEFO = []
cost_DEFO = []
cost_SAP = []
cost_DRL_SP = []
cost_DRL_HILL = []
cost_HILL = []
for elem in new_uti_ALL:
if elem[1]<-0.5:
uti_SAP.append(-0.5)
else:
uti_SAP.append(elem[1])
uti_DRL_SP.append(elem[2])
uti_DRL_HILL.append(elem[4])
uti_HILL.append(elem[8])
uti_DEFO.append(elem[10])
cost_SAP.append(elem[5])
cost_DRL_SP.append(elem[6])
cost_DRL_HILL.append(elem[7])
cost_HILL.append(elem[9])
cost_DEFO.append(elem[11])
plt.rcParams.update({'font.size': 16})
plt.rcParams['pdf.fonttype'] = 42
# Force ticks to appear
plt.yticks(np.arange(-50, 60, 10))
plt.ylim((-50, 60))
plt.xticks(np.arange(0, 75, 8))
plt.plot(X_axis, np.array(uti_SAP)*100, c='darkorange', linestyle='--', label="SAP", linewidth=3)
#plt.plot(X_axis, uti_MLP, c='aqua', linestyle='-', label="MLP", linewidth=3)
#plt.plot(X_axis, np.array(uti_DRL_SP)*100, c='dimgrey', linestyle='-.', label="DRL+GNN", linewidth=3)
plt.plot(X_axis, np.array(uti_HILL)*100, c='darkgreen', linestyle='-', label="LS", linewidth=3)
plt.plot(X_axis, np.array(uti_DEFO)*100, c='darkblue', linestyle=':', label="DEFO", linewidth=3)
plt.plot(X_axis, np.array(uti_DRL_HILL)*100, c='red', linestyle='-.', label="Enero", linewidth=3)
plt.ylabel("Relative performance w.r.t.\nLocal Search (%)", fontsize=16)
plt.xlabel("Topology identifier", fontsize=19)
#lgd = plt.legend(loc="lower left", bbox_to_anchor=(-0.3, -0.4), ncol=4)
plt.grid(c='grey')
plt.tight_layout()
plt.savefig(path_to_dir+"Figure_9_Rel_Perf_all_tops.pdf", bbox_inches='tight')
# plt.show()
plt.clf()
plt.close()
plt.rcParams.update({'font.size': 16})
plt.rcParams['pdf.fonttype'] = 42
# Force ticks to appear
plt.yticks(np.arange(0, 200, 20))
plt.ylim((0, 200))
plt.xticks(np.arange(0, 75, 8))
plt.plot(X_axis, np.array(cost_SAP), c='darkorange', linestyle='--', label="SAP", linewidth=2)
plt.plot(X_axis, np.array(cost_HILL), c='darkgreen', linestyle='-', label="LS", linewidth=2)
#plt.plot(X_axis, uti_MLP, c='aqua', linestyle='-', label="MLP", linewidth=2)
#plt.plot(X_axis, np.array(cost_DRL_SP), c='dimgrey', linestyle='-.', label="DRL+GNN", linewidth=2)
plt.plot(X_axis, np.array(cost_DEFO), c='darkblue', linestyle=':', label="DEFO", linewidth=2)
plt.plot(X_axis, np.array(cost_DRL_HILL), c='red', linestyle='-.', label="Enero", linewidth=2)
plt.ylabel("Average Execution Cost (s)", fontsize=16)
plt.xlabel("Topology identifier", fontsize=19)
plt.grid(c='grey')
lgd = plt.legend(loc="lower left", bbox_to_anchor=(-0.2, -0.33), ncol=4)
plt.savefig(path_to_dir+"Cost_all_tops.pdf", bbox_extra_artists=(lgd,), bbox_inches='tight')
# plt.show()
plt.close()
fig, ax = plt.subplots()
n = np.arange(1,len(cost_SAP)+1) / np.float(len(cost_SAP))
Xs = np.sort(cost_SAP)
ax.step(Xs,n, c='darkorange', linestyle='--', label="SAP", linewidth=3)
Xs = np.sort(cost_HILL)
ax.step(Xs,n,c='darkgreen', linestyle='-', label="LS", linewidth=3)
Xs = np.sort(cost_DEFO)
ax.step(Xs,n,c='darkblue', linestyle=':', label="DEFO", linewidth=3)
Xs = np.sort(cost_DRL_HILL)
ax.step(Xs,n,c='red', linestyle='-.', label="Enero", linewidth=3)
t = ax.text(50.0, 0.30, "Better", ha="center", va="center", size=15,
bbox=dict(boxstyle="larrow,pad=0.2", fc="w", ec="k", lw=2))
t = ax.text(165.0, 0.70, "DEFO", c="white", rotation=-45, ha="center", va="center", size=15,
bbox=dict(boxstyle="rarrow,pad=0.2", fc="blue", ec="w", lw=2))
plt.ylim((0, 1.005))
plt.xlim((-2, 185.0))
plt.xticks(np.arange(0, 185, 20))
plt.ylabel('CDF', fontsize=17)
plt.xlabel("Execution Cost (s)", fontsize=20)
plt.legend(prop={'size': 12, 'weight': 'bold'}, loc='lower right')
plt.grid(color='gray')
lgd = plt.legend(loc="lower left", bbox_to_anchor=(-0.2, -0.4), ncol=4)
plt.tight_layout()
#plt.show()
plt.savefig(path_to_dir+'Figure_9_CDF_Cost_AllTops.pdf', bbox_extra_artists=(lgd,), bbox_inches='tight')
#plt.clf()