def procedure1_nd_tree(n, p, k, w):
print("procedure1_ND_tree n={} p={}".format(n, p))
data = file_parsing.get_data(n, p)
# print("Recherche locale:\n\n")
time1 = time.time()
[allx, ally] = nd_tree.nd_tree(n, k, p, data)
time2 = time.time()
ally_for_file = copy.deepcopy(ally)
# print("Données utilisées: ", data)
# print("Vecteurs d'affectation solutions: ", allx)
# print("Valeurs des évaluations: ", ally)
# print("\n\nElicitation incrémentale:\nNombre de solutions potentielles:", len(ally), "\n\n")
time3 = time.time()
[opt, opt_value, nb_q, _,
_] = incremental_elicitation.mmr_incremental_elicitaiton(
allx, ally, w, [])
time4 = time.time()
# print("Solution optimale: ", opt)
# print("Valeur de la solution: ", opt_value)
# print("Poids du décideur: ", w)
# print("Nombre de questions: ", nb_q)
real_opt = compare_to_file(n, p, k)
file_parsing.write_res_proc1_nd_tree(opt_value, real_opt, n, p, w, nb_q,
time2 - time1, time4 - time3)
def procedure2_interactive_local_search(n, p, k, w, start_time):
print("START procedure2_ILS n={} p={}".format(n, p))
data = file_parsing.get_data(n, p) # n: nb de objets; p: nb de criteres
# print("Recherche locale + Elicitation incrementale:\n\n")
time1 = time.time()
opt, opt_solution, opt_value, nb_q = interactive_local_search.interactive_local_search(
n, k, p, data, w)
time2 = time.time()
#print("Solution optimale: ", opt)
# print("Valeur de la solution: ", opt_value)
# print("Poids du décideur: ", w)
# print("Nombre de questions: ", nb_q)
real_opt = compare_to_file(n, p, k)
real_opt = sum(w[i] * real_opt[i] for i in range(len(w)))
file_parsing.write_res_proc2_ILS(opt_value, real_opt, n, p, w, nb_q,
time2 - time1)
print("END procedure2_ILS n={} p={} time={}".format(
n, p,
time.time() - start_time))
return
def procedure1_PLS(n, p, k, w, start_time):
print("procedure1_PLS n={} p={}".format(n, p))
data = file_parsing.get_data(n, p) # n: nb de objets; p: nb de criteres
# print("Recherche locale:\n\n")
time1 = time.time()
[allx, ally] = local_search.neighbor_local_search(n, k, p, data)
time2 = time.time()
ally_for_file = copy.deepcopy(ally)
#print("Données utilisées: ", data)
#print("Vecteurs d'affectation solutions: ", allx)
#print("Valeurs des évaluations: ", ally)
# print("\n\nElicitation incrémentale:\nNombre de solutions potentielles:", len(ally), "\n\n")
time3 = time.time()
evidence = []
[opt, opty, opt_value, nb_q, _,
_] = incremental_elicitation.mmr_incremental_elicitaiton(
allx, ally, w, evidence)
time4 = time.time()
# print("Solution optimale: ", opt)
# print("Valeur de la solution: ", opt_value)
# print("Poids du décideur: ", w)
# print("Nombre de questions: ", nb_q)
real_opt = compare_to_file(n, p, k)
real_opt = sum(w[i] * real_opt[i] for i in range(len(w)))
file_parsing.write_res_proc1_PLS(opt_value, real_opt, n, p, w, nb_q,
time2 - time1, time4 - time3)
print("END procedure1_PLS n={} p={} time={}".format(n, p, time.time()))
return
time2 - time1)
print("END procedure2_ILS n={} p={} time={}".format(
n, p,
time.time() - start_time))
return
if __name__ == "__main__":
# n objets, entre 1 et 200
parametres = [[10, 2], [10, 3], [10, 4], [10, 5], [20, 3], [20, 4]]
para_ILS = [[20, 5], [30, 4], [30, 5], [40, 3], [40, 4]]
for n, p in para_ILS:
k = math.floor(n / 2)
data = file_parsing.get_data(n,
p) # n: nb de objets; p: nb de criteres
###################################################
##### MEME W POUR TROIS ALGO!
######################################################
# on crée un vecteur de poids aléatoire simulant les préférences du décideur
w = [random.uniform(0, 1) for i in range(p)]
# on le normalise pour que la somme du vecteur fasse 1
w = [w[i] / sum(w) for i in range(p)]
for i in range(4):
procedure2_interactive_local_search(n, p, k, w, time.time())
#procedure1_PLS(n,p,k,w,time.time());
# for i in range(4):
# procedure1_nd_tree(n,p,k,w,time.time());