class IterativeImprovement:
def __init__(self):
self.test_data_path = './PFSP_benchmark_data_set/tai50_20_1.txt'
self.tool = Tool()
self.max_span_time = 10000
self.span = self.tool.io(self.test_data_path) # 測資
self.job_len = len(self.span[0])
self.eval_time = 20
def generateNeighborList(self, jobs_seq):
ret = []
ret.append(jobs_seq)
for i in range(len(jobs_seq)):
for j in range(i, len(jobs_seq)):
if (i != j):
tmp_seq = copy.deepcopy(jobs_seq)
tmp_seq[i], tmp_seq[j] = tmp_seq[j], tmp_seq[i]
ret.append(tmp_seq)
return ret
def run(self, jobs_seq):
min_span = self.tool.makespan(self.span, jobs_seq) # 起始min值
span_list = []
count = 0
while (count < self.max_span_time):
nb_list = self.generateNeighborList(jobs_seq)
for i in range(len(nb_list)):
span = self.tool.makespan(self.span, nb_list[i])
count += 1
if (span < min_span):
min_span = span
jobs_seq = nb_list[i]
if (jobs_seq == nb_list[0]):
return min_span
span_list.append(min_span)
return min_span
def average(self, lst):
return sum(lst) / len(lst)
def search(self):
pool = mp.Pool(mp.cpu_count())
random.seed(0)
init_list = []
for i in range(self.eval_time):
jobs_seq = [int(e) for e in range(0, self.job_len)]
random.shuffle(jobs_seq)
init_list.append(jobs_seq)
pool_out = pool.map_async(self.run, init_list)
pool.close()
pool.join()
result_list = pool_out.get()
print(result_list)
print('min', min(result_list))
print('max', max(result_list))
print('avg', self.average(result_list))
