-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathpart3.py
144 lines (112 loc) · 5.63 KB
/
part3.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
import gurobipy as grb
import numpy as np
from utils import *
import pandas as pd
import sys
import time
def generate_knapsack_instance(n, p, utility_range=(0, 20), weight_range=(0, 10)):
objects_utilities = np.random.randint(utility_range[0], utility_range[1], size=(p, n))
objects_weights = np.random.randint(weight_range[0], weight_range[1], size=p)
capacity = np.sum(objects_weights) / 2
return objects_utilities, objects_weights, capacity
def knapsack_plmo(instance, ideal, nadir):
model = grb.Model("PLMO Knapsack (" + str(instance[0].shape[0]) + ", " + str(instance[1].shape[0]) + ")")
z_variable = model.addVar(name="z")
obj_variables = np.array([_ for _ in model.addVars(instance[1].shape[0], vtype=grb.GRB.BINARY, name="w").values()])
utility_expr = (ideal - instance[0].transpose().dot(obj_variables)) / (ideal - nadir + sys.float_info.epsilon)
aggregated_utilities = np.sum(utility_expr)
model.setObjective(z_variable + aggregated_utilities, grb.GRB.MINIMIZE)
for idx, criterion in enumerate(utility_expr):
model.addConstr(z_variable >= criterion, "u" + str(idx))
model.addConstr(np.dot(instance[1], obj_variables) <= instance[2])
model.setParam('OutputFlag', False)
model.optimize()
model.write("knapsack_plmo.lp")
return model.objVal, [v.x for v in model.getVars()][1:]
def interaction(knapsack):
processed_data = {str(i) : d for (i, d) in enumerate(knapsack[0])}
raw_data = pd.DataFrame.from_dict(processed_data, orient='index', columns=["u" + str(i) + "(max)" for i in range(knapsack[0].shape[1])])
print(raw_data)
stop = False
pareto_front = get_pareto(raw_data)
print("pareto :", pareto_front)
while not stop:
# process Nadir
nadir = np.array([_ for _ in get_nadir(raw_data, pareto_front).values()])
ideal = np.array([_ for _ in get_ideal(raw_data, pareto_front).values()])
print("***new nadir \n",nadir)
value, elems = knapsack_plmo(knapsack, ideal, nadir)
#values, best_index = tchebycheff_augmente(data, pareto, w, ideal, nadir)
#print("Solution qu'on vous propose : ", df['nom'][best_index],"\nValeurs :")
# show result
for idx, elem in enumerate(elems):
if np.isclose(elem, 1):
print("object : {}".format(idx))
for criterion in raw_data.columns:
if np.isclose(elem, 1):
#print(raw_data[criterion])
print(criterion, ":", raw_data[criterion][idx])
# ask if satisfied
satisfait = input("Etes-vous satisfait de la solution retournée? \no : oui \nn : non \nvotre réponse : ")
if satisfait == "o" :
stop = True
else :
# get favorite criterion to improve
print("************\nListe des critères : ",raw_data.columns.values)
c = input("Quel est le critère que vous voulez favoriser ?")
while c not in raw_data.columns:
c = input("Quel est le critère que vous voulez favoriser ?")
bound = np.min(raw_data[c])
print("---- criterion : ", c," bound =", bound)
# reduce pareto front
pareto_front = update_pareto(raw_data, pareto_front, c, bound)
# check if there are solutions left
pareto_list = get_paretoList(pareto_front)
if len(pareto_list) == 0:
print("Aucune solution ne correspond à vos critères")
stop = True
# print("***new pareto \n",pareto)
def benchmark():
nb_objects_steps = [1, 5, 10, 20, 50, 100, 200, 500, 1000]
nb_criterion_steps = [1, 5, 10, 20, 50, 100, 200, 500, 1000]
benchmark_times = []
for idx1, s1 in enumerate(nb_objects_steps):
print("objects : ", s1)
benchmark_times.append([])
for idx2, s2 in enumerate(nb_criterion_steps):
print("criterion : ", s2)
knapsack = generate_knapsack_instance(s2, s1)
data = {str(i) : d for (i, d) in enumerate(knapsack[0])}
raw_data = pd.DataFrame.from_dict(data, orient='index', columns=["u" + str(i) + "(max)" for i in range(knapsack[0].shape[1])])
pareto_front = get_pareto(raw_data)
nadir = np.array([_ for _ in get_nadir(raw_data, pareto_front).values()])
ideal = np.array([_ for _ in get_ideal(raw_data, pareto_front).values()])
current_time = 0
r = 100
if s1 >= 500 and s2 >= 500:
r = 1
for i in range(r):
start = time.time()
knapsack_plmo(knapsack, ideal, nadir)
end = time.time()
current_time += (end - start)
benchmark_times[-1].append(current_time / r)
print(benchmark_times)
if __name__ == "__main__":
nb_criterion = 2
nb_objects = 10
knapsack = generate_knapsack_instance(nb_criterion, nb_objects)
#res = knapsack_plmo(knapsack)
data = {str(i) : d for (i, d) in enumerate(knapsack[0])}
raw_data = pd.DataFrame.from_dict(data, orient='index', columns=["u" + str(i) + "(max)" for i in range(knapsack[0].shape[1])])
pareto_front = get_pareto(raw_data)
nadir = np.array([_ for _ in get_nadir(raw_data, pareto_front).values()])
ideal = np.array([_ for _ in get_ideal(raw_data, pareto_front).values()])
print(nadir, ideal)
res = knapsack_plmo(knapsack, ideal, nadir)
print(res)
print("Capacity : ", knapsack[2])
print("Items total weight : ", np.dot(np.array(res[1]), knapsack[1]))
print("Items total utility : ", np.sum(np.dot(np.array(res[1]), knapsack[0])))
benchmark()
# interaction(knapsack)