def vns(optimum_value: float, instance_dict: dict, output_filename: str) -> Solution:
config = settings.EVALUATE_METHODS_SETTINGS
max_iterations = config.get('vns', {}).get('max_iterations')
neighborhood_size = config.get('vns', {}).get('neighborhood_size')
solution = Solution(
n=instance_dict.get('n'),
capacity=instance_dict.get('capacity'),
optimum=optimum_value
)
solution.generate_starter_solution(
item_list=instance_dict.get('item_list'),
random_seed=RANDOM_SEED
)
#solution.print_solution()
vns_method = Vns(
item_list=instance_dict.get('item_list'),
neighborhood_size=neighborhood_size
)
vns_method.run_vns(
solution=solution,
item_list=instance_dict.get('item_list'),
max_iterations=max_iterations,
neighborhood_size=neighborhood_size,
output_filename=output_filename
)
print(f"ic| VNS Optimum Solution")
solution.print_solution(item_list=instance_dict.get('item_list'))
return solution
def simulated_annealing(optimum_value: float, instance_dict: dict, output_filename: str):
config = settings.EVALUATE_METHODS_SETTINGS
max_iterations = config.get('sa', {}).get('max_iterations')
distance = config.get('sa', {}).get('distance')
initial_temperature = config.get('sa', {}).get('initial_temperature')
best_solution = Solution(
n=instance_dict.get('n'),
capacity=instance_dict.get('capacity'),
optimum=optimum_value
)
best_solution.generate_starter_solution(
item_list=instance_dict.get('item_list'),
random_seed=RANDOM_SEED
)
best_solution.print_solution()
sa = SimulatedAnnealing(
solution=deepcopy(best_solution),
item_list=instance_dict.get('item_list'),
distance=2,
output_filename=f"{output_filename}_temp"
)
current_solution = deepcopy(sa.solution)
writer = FileWriter(file_name=f"{output_filename}_temp")
for i in range(max_iterations):
if i == 0:
writer.write_line(output_filename.replace('TEMP-', ''))
writer.write_line(str(current_solution.optimum))
writer.write_line(f"{i} {current_solution.value}")
sa.solution = deepcopy(current_solution)
# random neighbor
test = sa.random_neighbor(distance=distance)
if not test:
print(f"ERROR ! Could not find a random neighbor")
# if better set as best_solution
if sa.solution.value > best_solution.value:
best_solution = deepcopy(sa.solution)
writer.write_line(f"{i} {best_solution.value}")
# calculate diff
diff = current_solution.value - sa.solution.value
# calculate temp
t = initial_temperature / float(i + 1)
if diff < 0 or random() < exp(-diff / t):
current_solution = deepcopy(sa.solution)
print(f"ic| SA Optimum Solution")
best_solution.print_solution(item_list=instance_dict.get('item_list'))
return best_solution
def iterated_local_search(optimum_value: float, instance_dict: dict,
output_filename: str) -> Solution:
config = settings.EVALUATE_METHODS_SETTINGS
max_iterations = config.get('ils', {}).get('max_iterations')
best_solution = Solution(n=instance_dict.get('n'),
capacity=instance_dict.get('capacity'),
optimum=optimum_value)
best_solution.generate_starter_solution(
item_list=instance_dict.get('item_list'), random_seed=RANDOM_SEED)
counter = 0
for i in range(max_iterations):
perturbed_solution = deepcopy(best_solution)
perturbed_solution.perturb_solution(
item_list=instance_dict.get('item_list'))
print(f"Iteration {i}")
if i == 0:
ils = IteratedLocalSearch(
solution=perturbed_solution,
item_list=instance_dict.get('item_list'),
distance=2,
output_filename=f"{output_filename}_temp")
else:
ils.solution = deepcopy(perturbed_solution)
ils.run(counter)
while not ils.check_acceptance_criteria(best_solution):
ils.solution.perturb_solution(
item_list=instance_dict.get('item_list'))
ils.run(counter)
if ils.solution.value > best_solution.value:
counter += 1
best_solution = deepcopy(ils.solution)
best_solution.print_solution()
ils.output_file.write_line(f"{counter} {ils.solution.value}")
print(f"ic| ILS Optimum Solution")
best_solution.print_solution(item_list=instance_dict.get('item_list'))
return best_solution
def single_start_local_search(optimum_value: float, instance_dict: dict,
output_filename: str) -> Solution:
solution = Solution(n=instance_dict.get('n'),
capacity=instance_dict.get('capacity'),
optimum=optimum_value)
solution.generate_starter_solution(
item_list=instance_dict.get('item_list'), random_seed=RANDOM_SEED)
#solution.print_solution()
local_search = LocalSearch(solution=solution,
item_list=instance_dict.get('item_list'),
distance=2,
output_filename=output_filename,
improved=True)
local_search.run()
print(f"ic| SSL Optimum Solution")
solution.print_solution(item_list=instance_dict.get('item_list'))
return solution
def tabu_search(optimum_value: float, instance_dict: dict, output_filename: str):
config = settings.EVALUATE_METHODS_SETTINGS
max_iterations = config.get('tabu', {}).get('max_iterations')
best_solution = Solution(
n=instance_dict.get('n'),
capacity=instance_dict.get('capacity'),
optimum=optimum_value
)
best_solution.generate_starter_solution(
item_list=instance_dict.get('item_list'),
random_seed=RANDOM_SEED
)
best_solution.print_solution()
tabu_list = {}
tabu_tenure = config.get('tabu', {}).get('tenure')
tabu_search = TabuSearch(
solution=deepcopy(best_solution),
item_list=instance_dict.get('item_list'),
distance=2,
output_filename=f"{output_filename}_temp"
)
for _ in range(max_iterations):
tabu_search.run(tabu_list=tabu_list)
if tabu_search.solution.value > best_solution.value:
best_solution = deepcopy(tabu_search.solution)
# decrement tenure, removes if 0
for solution in tabu_list.copy():
tabu_list[solution] -= 1
if tabu_list[solution] == 0:
del tabu_list[solution]
# add new solution to list
number = int(''.join([str(x) for x in tabu_search.solution.item_list]), 2)
tabu_list[number] = tabu_tenure
print(f"ic| Tabu Optimum Solution")
best_solution.print_solution(item_list=instance_dict.get('item_list'))
return best_solution
def multi_start_local_search(optimum_value: float, instance_dict: dict,
output_filename: str) -> Solution:
config = settings.EVALUATE_METHODS_SETTINGS
max_iterations = config.get('msl', {}).get('max_iterations')
best_solution = None
counter = 0
msl = None
for i in range(max_iterations):
print(f"Iteration {i}")
random_solution = Solution(n=instance_dict.get('n'),
capacity=instance_dict.get('capacity'),
optimum=optimum_value)
random_solution.generate_starter_solution(
item_list=instance_dict.get('item_list'),
random_seed=(RANDOM_SEED + i))
#random_solution.print_solution()
if i == 0:
msl = MultiStartLocalSearch(
solution=random_solution,
item_list=instance_dict.get('item_list'),
distance=2,
output_filename=f"{output_filename}_temp")
else:
msl.solution = deepcopy(random_solution)
msl.run(counter=counter)
if best_solution is None or msl.solution.value > best_solution.value:
counter += 1
msl.output_file.write_line(f"{counter} {msl.solution.value}")
best_solution = deepcopy(msl.solution)
print(f"ic| MSL Optimum Solution")
best_solution.print_solution(item_list=instance_dict.get('item_list'))
return best_solution
