-
Notifications
You must be signed in to change notification settings - Fork 0
/
algorithms.py
executable file
·182 lines (134 loc) · 4.27 KB
/
algorithms.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
def baseline_greedy_uncons(obs_util, grps, k, m, pr=False):
n = np.sum([len(g) for g in grps])
all = [i for i in range(n)]
sol = [] #()
for i in range(k):
utils = marg_F_mult(sol, all, obs_util, m)
tmp_sol = get_top_n(all, utils, 1)
assert(len(tmp_sol) == 1)
ind = list(tmp_sol)[0]
all.remove(ind)
sol.append(ind)
assert(len(sol)==k)
return sol
def baseline_greedy_rooney(obs_util, grps, k, m, pr=False):
n = np.sum([len(g) for g in grps])
all = [i for i in range(n)]
sol = [] #()
for i in range(k):
utils = marg_F_mult(sol, all, obs_util, m)
if i == k-1:
disadv_selected = False
for s in sol:
if s in grps[1]:
disadv_selected = True
break
if not disadv_selected:
all = list(grps[1])
tmp_sol = get_top_n(all, utils, 1)
assert(len(tmp_sol) == 1)
ind = list(tmp_sol)[0]
all.remove(ind)
sol.append(ind)
assert(len(sol)==k)
return sol
def baseline_greedy_equal(obs_util, grps, k, m):
p = len(grps)
n = np.sum([len(g) for g in grps])
all = [i for i in range(n)]
pr = False
sol = [] #list()
cnt = [0 for t in range(p)]
while len(sol) < k:
utils = marg_F_mult(sol, all, obs_util, m)
tmp_sol = get_top_n(all, utils, 1)
assert(len(tmp_sol) == 1)
ind = list(tmp_sol)[0]
inv_ind = -1
for i in range(len(all)):
if all[i] == ind:
inv_ind = i
break
fg = 1
for t in range(p):
if ind in grps[t]:
if cnt[t] > max(k * 0.5, k - len(grps[1-t])):
fg = 0
if fg:
sol.append(ind)
for t in range(p):
if ind in grps[t]:
cnt[t] += 1
all.remove(ind)
assert(len(sol)==k)
return sol
def baseline_greedy_cons(obs_util, grps, k, m):
p = len(grps)
n = np.sum([len(g) for g in grps])
all = [i for i in range(n)]
pr = False
sol = [] #list()
cnt = [0 for t in range(p)]
while len(sol) < k:
utils = marg_F_mult(sol, all, obs_util, m)
tmp_sol = get_top_n(all, utils, 1)
assert(len(tmp_sol) == 1)
ind = list(tmp_sol)[0]
inv_ind = -1
for i in range(len(all)):
if all[i] == ind:
inv_ind = i
break
fg = 1
for t in range(p):
if ind in grps[t]:
if cnt[t] >= k * len(grps[t]) / n:
fg = 0
if fg:
sol.append(ind)
for t in range(p):
if ind in grps[t]:
cnt[t] += 1
all.remove(ind)
assert(len(sol)==k)
return sol
def algo_3_disjoint_attr(obs_util, grps, k, m): # lat_util = None):
p = len(grps)
n = np.sum([len(g) for g in grps])
grp_lists = [list(grps[t]) for t in range(p)]
def count_sol_in_attr(sol, obs_util):
k_pr = np.array([0.0 for j in range(m)])
for j in range(m):
for i in sol_tmp:
k_pr[j] += (obs_util[i][j] > 0)
k_pr *= float(float(k)/np.sum(k_pr))
return k_pr
#####################################
obs_util_grp_a = copy.deepcopy(obs_util)
for i in grps[1]: obs_util_grp_a[i] *= 0
sol_tmp = baseline_greedy_uncons(obs_util_grp_a, grps, k * len(grps[0]) // n, m, pr=False)
k_pr = count_sol_in_attr(sol_tmp, obs_util)
####################################
sol = []
for j in range(m):
for t in range(p):
k_int = int(k_pr[j] * len(grps[t]) / n)
utils = [obs_util[i][j] for i in grps[t]]
tmp_sol = get_top_n(grps[t], utils, k_int)
for tmp in tmp_sol:
sol.append(tmp)
sol = set(sol)
if len(sol) < k:
unselec = []
for i in range(n):
if i not in sol:
unselec.append(i)
utils = marg_F_mult(sol, unselec, obs_util, m)
tmp_sol = get_top_n(unselec, utils, k - len(sol))
for i in tmp_sol:
sol.add(i)
if len(sol) == k:
break
sol = list(sol)
assert(len(sol)==k)
return sol