def get_result(self) -> List[FloatSolution]:
archive = NonDominatedSolutionListArchive()
for solution in self.solutions:
archive.add(solution)
return archive.solution_list
def __init__(self, problem: Problem[S],
termination_criterion: TerminationCriterion):
super().__init__()
self.problem = problem
self.termination_criterion = termination_criterion
self.observable.register(termination_criterion)
self.archive = NonDominatedSolutionListArchive()
class RandomSearch(Algorithm[S, R]):
def __init__(self,
problem: Problem[S],
termination_criterion: TerminationCriterion = store.
default_termination_criteria):
super().__init__()
self.problem = problem
self.termination_criterion = termination_criterion
self.observable.register(termination_criterion)
self.archive = NonDominatedSolutionListArchive()
def get_observable_data(self) -> dict:
ctime = time.time() - self.start_computing_time
return {
'PROBLEM': self.problem,
'EVALUATIONS': self.evaluations,
'SOLUTIONS': self.get_result(),
'COMPUTING_TIME': ctime
}
def create_initial_solutions(self) -> List[S]:
return [self.problem.create_solution()]
def evaluate(self, solution_list: List[S]) -> List[S]:
return [self.problem.evaluate(solution_list[0])]
def init_progress(self) -> None:
self.evaluations = 1
observable_data = self.get_observable_data()
self.observable.notify_all(**observable_data)
def stopping_condition_is_met(self) -> bool:
return self.termination_criterion.is_met
def step(self) -> None:
new_solution = self.problem.create_solution()
self.problem.evaluate(new_solution)
self.archive.add(new_solution)
def update_progress(self) -> None:
self.evaluations += 1
observable_data = self.get_observable_data()
self.observable.notify_all(**observable_data)
def get_result(self) -> List[S]:
return self.archive.solution_list
def get_name(self) -> str:
return 'Random Search'
@property
def label(self) -> str:
return f'{self.get_name()}.{self.problem.get_name()}'
def __init__(self,
problem: FloatProblem,
swarm_size: int,
uniform_mutation: UniformMutation,
non_uniform_mutation: NonUniformMutation,
leaders: Optional[BoundedArchive],
epsilon: float,
termination_criterion: TerminationCriterion,
swarm_generator: Generator = store.default_generator,
swarm_evaluator: Evaluator = store.default_evaluator):
""" This class implements the OMOPSO algorithm as described in
todo Update this reference
* SMPSO: A new PSO-based metaheuristic for multi-objective optimization
The implementation of OMOPSO provided in jMetalPy follows the algorithm template described in the algorithm
templates section of the documentation.
:param problem: The problem to solve.
:param swarm_size: Size of the swarm.
:param leaders: Archive for leaders.
"""
super(OMOPSO, self).__init__(problem=problem, swarm_size=swarm_size)
self.swarm_generator = swarm_generator
self.swarm_evaluator = swarm_evaluator
self.termination_criterion = termination_criterion
self.observable.register(termination_criterion)
self.uniform_mutation = uniform_mutation
self.non_uniform_mutation = non_uniform_mutation
self.leaders = leaders
self.epsilon = epsilon
self.epsilon_archive = NonDominatedSolutionListArchive(
EpsilonDominanceComparator(epsilon))
self.c1_min = 1.5
self.c1_max = 2.0
self.c2_min = 1.5
self.c2_max = 2.0
self.r1_min = 0.0
self.r1_max = 1.0
self.r2_min = 0.0
self.r2_max = 1.0
self.weight_min = 0.1
self.weight_max = 0.5
self.change_velocity1 = -1
self.change_velocity2 = -1
self.dominance_comparator = DominanceComparator()
self.speed = numpy.zeros(
(self.swarm_size, self.problem.number_of_variables), dtype=float)
def __init__(
self,
problem: BinaryProblem,
swarm_size: int,
mutation: BitFlipMutation,
leaders: Optional[BoundedArchive],
epsilon: float,
termination_criterion: TerminationCriterion,
swarm_generator: Generator = store.default_generator,
swarm_evaluator: Evaluator = store.default_evaluator,
):
super(BMOPSO, self).__init__(problem=problem, swarm_size=swarm_size)
self.swarm_generator = swarm_generator
self.swarm_evaluator = swarm_evaluator
self.termination_criterion = termination_criterion
self.observable.register(termination_criterion)
self.mutation_operator = mutation
self.leaders = leaders
self.epsilon = epsilon
self.epsilon_archive = NonDominatedSolutionListArchive(
# EpsilonDominanceComparator(epsilon)
DominanceComparator())
self.c1_min = 1.5
self.c1_max = 2.0
self.c2_min = 1.5
self.c2_max = 2.0
self.r1_min = 0.0
self.r1_max = 1.0
self.r2_min = 0.0
self.r2_max = 1.0
self.weight_min = 0.1
self.weight_max = 0.5
self.change_velocity1 = -1
self.change_velocity2 = -1
self.dominance_comparator = DominanceComparator()
self.speed = numpy.zeros(
(
self.swarm_size,
self.problem.number_of_variables,
self.problem.number_of_tests,
),
dtype=float,
)
def get_non_dominated_solutions(solutions: List[Solution]) -> List[Solution]:
archive: BoundedArchive = NonDominatedSolutionListArchive()
for solution in solutions:
archive.add(solution)
return archive.solution_list
class OMOPSO(ParticleSwarmOptimization):
def __init__(self,
problem: FloatProblem,
swarm_size: int,
uniform_mutation: UniformMutation,
non_uniform_mutation: NonUniformMutation,
leaders: Optional[BoundedArchive],
epsilon: float,
termination_criterion: TerminationCriterion,
swarm_generator: Generator = store.default_generator,
swarm_evaluator: Evaluator = store.default_evaluator):
""" This class implements the OMOPSO algorithm as described in
todo Update this reference
* SMPSO: A new PSO-based metaheuristic for multi-objective optimization
The implementation of OMOPSO provided in jMetalPy follows the algorithm template described in the algorithm
templates section of the documentation.
:param problem: The problem to solve.
:param swarm_size: Size of the swarm.
:param leaders: Archive for leaders.
"""
super(OMOPSO, self).__init__(problem=problem, swarm_size=swarm_size)
self.swarm_generator = swarm_generator
self.swarm_evaluator = swarm_evaluator
self.termination_criterion = termination_criterion
self.observable.register(termination_criterion)
self.uniform_mutation = uniform_mutation
self.non_uniform_mutation = non_uniform_mutation
self.leaders = leaders
self.epsilon = epsilon
self.epsilon_archive = NonDominatedSolutionListArchive(
EpsilonDominanceComparator(epsilon))
self.c1_min = 1.5
self.c1_max = 2.0
self.c2_min = 1.5
self.c2_max = 2.0
self.r1_min = 0.0
self.r1_max = 1.0
self.r2_min = 0.0
self.r2_max = 1.0
self.weight_min = 0.1
self.weight_max = 0.5
self.change_velocity1 = -1
self.change_velocity2 = -1
self.dominance_comparator = DominanceComparator()
self.speed = numpy.zeros(
(self.swarm_size, self.problem.number_of_variables), dtype=float)
def create_initial_solutions(self) -> List[FloatSolution]:
return [
self.swarm_generator.new(self.problem)
for _ in range(self.swarm_size)
]
def evaluate(self, solution_list: List[FloatSolution]):
return self.swarm_evaluator.evaluate(solution_list, self.problem)
def stopping_condition_is_met(self) -> bool:
return self.termination_criterion.is_met
def initialize_global_best(self, swarm: List[FloatSolution]) -> None:
for particle in swarm:
if self.leaders.add(particle):
self.epsilon_archive.add(copy(particle))
def initialize_particle_best(self, swarm: List[FloatSolution]) -> None:
for particle in swarm:
particle.attributes['local_best'] = copy(particle)
def initialize_velocity(self, swarm: List[FloatSolution]) -> None:
for i in range(self.swarm_size):
for j in range(self.problem.number_of_variables):
self.speed[i][j] = 0.0
def update_velocity(self, swarm: List[FloatSolution]) -> None:
for i in range(self.swarm_size):
best_particle = copy(swarm[i].attributes['local_best'])
best_global = self.select_global_best()
r1 = round(random.uniform(self.r1_min, self.r1_max), 1)
r2 = round(random.uniform(self.r2_min, self.r2_max), 1)
c1 = round(random.uniform(self.c1_min, self.c1_max), 1)
c2 = round(random.uniform(self.c2_min, self.c2_max), 1)
w = round(random.uniform(self.weight_min, self.weight_max), 1)
for var in range(swarm[i].number_of_variables):
self.speed[i][var] = w * self.speed[i][var] \
+ (c1 * r1 * (best_particle.variables[var] - swarm[i].variables[var])) \
+ (c2 * r2 * (best_global.variables[var] - swarm[i].variables[var]))
def update_position(self, swarm: List[FloatSolution]) -> None:
for i in range(self.swarm_size):
particle = swarm[i]
for j in range(particle.number_of_variables):
particle.variables[j] += self.speed[i][j]
if particle.variables[j] < self.problem.lower_bound[j]:
particle.variables[j] = self.problem.lower_bound[j]
self.speed[i][j] *= self.change_velocity1
if particle.variables[j] > self.problem.upper_bound[j]:
particle.variables[j] = self.problem.upper_bound[j]
self.speed[i][j] *= self.change_velocity2
def update_global_best(self, swarm: List[FloatSolution]) -> None:
for particle in swarm:
if self.leaders.add(copy(particle)):
self.epsilon_archive.add(copy(particle))
def update_particle_best(self, swarm: List[FloatSolution]) -> None:
for i in range(self.swarm_size):
flag = self.dominance_comparator.compare(
swarm[i], swarm[i].attributes['local_best'])
if flag != 1:
swarm[i].attributes['local_best'] = copy(swarm[i])
def perturbation(self, swarm: List[FloatSolution]) -> None:
self.non_uniform_mutation.set_current_iteration(self.evaluations /
self.swarm_size)
for i in range(self.swarm_size):
if (i % 3) == 0:
self.non_uniform_mutation.execute(swarm[i])
else:
self.uniform_mutation.execute(swarm[i])
def select_global_best(self) -> FloatSolution:
leaders = self.leaders.solution_list
if len(leaders) > 2:
particles = random.sample(leaders, 2)
if self.leaders.comparator.compare(particles[0], particles[1]) < 1:
best_global = copy(particles[0])
else:
best_global = copy(particles[1])
else:
best_global = copy(self.leaders.solution_list[0])
return best_global
def __velocity_constriction(self, value: float, delta_max: [],
delta_min: [], variable_index: int) -> float:
result = value
if value > delta_max[variable_index]:
result = delta_max[variable_index]
if value < delta_min[variable_index]:
result = delta_min[variable_index]
return result
def __inertia_weight(self, wmax: float):
return wmax
def __constriction_coefficient(self, c1: float, c2: float) -> float:
rho = c1 + c2
if rho <= 4:
result = 1.0
else:
result = 2.0 / (2.0 - rho - sqrt(pow(rho, 2.0) - 4.0 * rho))
return result
def init_progress(self) -> None:
self.evaluations = self.swarm_size
self.leaders.compute_density_estimator()
self.initialize_velocity(self.solutions)
self.initialize_particle_best(self.solutions)
self.initialize_global_best(self.solutions)
def update_progress(self) -> None:
self.evaluations += self.swarm_size
self.leaders.compute_density_estimator()
observable_data = self.get_observable_data()
observable_data['SOLUTIONS'] = self.epsilon_archive.solution_list
self.observable.notify_all(**observable_data)
def get_result(self) -> List[FloatSolution]:
return self.epsilon_archive.solution_list
def get_name(self) -> str:
return 'OMOPSO'
def setUp(self):
self.archive = NonDominatedSolutionListArchive()
class NonDominatedSolutionListArchiveTestCases(unittest.TestCase):
def setUp(self):
self.archive = NonDominatedSolutionListArchive()
def test_should_constructor_create_a_non_null_object(self):
self.assertIsNotNone(self.archive)
def test_should_adding_one_solution_work_properly(self):
solution = Solution(1, 1)
self.archive.add(solution)
self.assertEqual(1, self.archive.size())
self.assertEqual(solution, self.archive.solution_list[0])
def test_should_adding_two_solutions_work_properly_if_one_is_dominated(self):
dominated_solution = Solution(1, 2)
dominated_solution.objectives = [2.0, 2.0]
dominant_solution = Solution(1, 2)
dominant_solution.objectives = [1.0, 1.0]
self.archive.add(dominated_solution)
self.archive.add(dominant_solution)
self.assertEqual(1, self.archive.size())
self.assertEqual(dominant_solution, self.archive.solution_list[0])
def test_should_adding_two_solutions_work_properly_if_both_are_non_dominated(self):
solution1 = Solution(1, 2)
solution1.objectives = [1.0, 0.0]
solution2 = Solution(1, 2)
solution2.objectives = [0.0, 1.0]
self.archive.add(solution1)
self.archive.add(solution2)
self.assertEqual(2, self.archive.size())
self.assertTrue(solution1 in self.archive.solution_list and
solution2 in self.archive.solution_list)
def test_should_adding_four_solutions_work_properly_if_one_dominates_the_others(self):
solution1 = Solution(1, 2)
solution1.objectives = [1.0, 1.0]
solution2 = Solution(1, 2)
solution2.objectives = [0.0, 2.0]
solution3 = Solution(1, 2)
solution3.objectives = [0.5, 1.5]
solution4 = Solution(1, 2)
solution4.objectives = [0.0, 0.0]
self.archive.add(solution1)
self.archive.add(solution2)
self.archive.add(solution3)
self.archive.add(solution4)
self.assertEqual(1, self.archive.size())
self.assertEqual(solution4, self.archive.solution_list[0])
def test_should_adding_three_solutions_work_properly_if_two_of_them_are_equal(self):
solution1 = Solution(1, 2)
solution1.objectives = [1.0, 1.0]
solution2 = Solution(1, 2)
solution2.objectives = [0.0, 2.0]
solution3 = Solution(1, 2)
solution3.objectives = [1.0, 1.0]
self.archive.add(solution1)
self.archive.add(solution2)
result = self.archive.add(solution3)
self.assertEqual(2, self.archive.size())
self.assertFalse(result)
self.assertTrue(solution1 in self.archive.solution_list
or solution3 in self.archive.solution_list)
class BMOPSO(ParticleSwarmOptimization):
def __init__(
self,
problem: BinaryProblem,
swarm_size: int,
mutation: BitFlipMutation,
leaders: Optional[BoundedArchive],
epsilon: float,
termination_criterion: TerminationCriterion,
swarm_generator: Generator = store.default_generator,
swarm_evaluator: Evaluator = store.default_evaluator,
):
super(BMOPSO, self).__init__(problem=problem, swarm_size=swarm_size)
self.swarm_generator = swarm_generator
self.swarm_evaluator = swarm_evaluator
self.termination_criterion = termination_criterion
self.observable.register(termination_criterion)
self.mutation_operator = mutation
self.leaders = leaders
self.epsilon = epsilon
self.epsilon_archive = NonDominatedSolutionListArchive(
# EpsilonDominanceComparator(epsilon)
DominanceComparator())
self.c1_min = 1.5
self.c1_max = 2.0
self.c2_min = 1.5
self.c2_max = 2.0
self.r1_min = 0.0
self.r1_max = 1.0
self.r2_min = 0.0
self.r2_max = 1.0
self.weight_min = 0.1
self.weight_max = 0.5
self.change_velocity1 = -1
self.change_velocity2 = -1
self.dominance_comparator = DominanceComparator()
self.speed = numpy.zeros(
(
self.swarm_size,
self.problem.number_of_variables,
self.problem.number_of_tests,
),
dtype=float,
)
def create_initial_solutions(self) -> List[BinarySolution]:
return [
self.swarm_generator.new(self.problem)
for _ in range(self.swarm_size)
]
def evaluate(self, solution_list: List[BinarySolution]):
return self.swarm_evaluator.evaluate(solution_list, self.problem)
def stopping_condition_is_met(self) -> bool:
return self.termination_criterion.is_met
def initialize_global_best(self, swarm: List[BinarySolution]) -> None:
for particle in swarm:
if self.leaders.add(particle):
self.epsilon_archive.add(copy(particle))
def initialize_particle_best(self, swarm: List[BinarySolution]) -> None:
for particle in swarm:
particle.attributes["local_best"] = copy(particle)
def initialize_velocity(self, swarm: List[BinarySolution]) -> None:
for i in range(self.swarm_size):
for j in range(self.problem.number_of_variables):
self.speed[i][j] = 0.0
def update_velocity(self, swarm: List[BinarySolution]) -> None:
for i in range(self.swarm_size):
best_particle = copy(swarm[i].attributes["local_best"])
best_global = self.select_global_best()
r1 = round(random.uniform(self.r1_min, self.r1_max), 1)
r2 = round(random.uniform(self.r2_min, self.r2_max), 1)
c1 = round(random.uniform(self.c1_min, self.c1_max), 1)
c2 = round(random.uniform(self.c2_min, self.c2_max), 1)
w = round(random.uniform(self.weight_min, self.weight_max), 1)
for var in range(swarm[i].number_of_variables):
best_particle_diff = numpy.subtract(
numpy.array(best_particle.variables[var]),
numpy.array(swarm[i].variables[var]),
dtype=numpy.float32,
)
best_global_diff = numpy.subtract(
numpy.array(best_global.variables[var]),
numpy.array(swarm[i].variables[var]),
dtype=numpy.float32,
)
self.speed[i][var] = (w * numpy.array(self.speed[i][var]) +
(c1 * r1 * best_particle_diff) +
(c2 * r2 * best_global_diff))
def update_position(self, swarm: List[BinarySolution]) -> None:
for i in range(self.swarm_size):
particle = swarm[i]
for j in range(particle.number_of_variables):
particle.variables[j] = self.compute_position(self.speed[i][j])
def compute_position(self, speed):
updated_positions = (numpy.random.random_sample(speed.shape) <
self._sigmoid(speed)) * 1
return list(numpy.array(updated_positions, dtype=bool))
def _sigmoid(self, x):
return 1 / (1 + numpy.exp(-x))
def update_global_best(self, swarm: List[BinarySolution]) -> None:
for particle in swarm:
if self.leaders.add(copy(particle)):
self.epsilon_archive.add(copy(particle))
def update_particle_best(self, swarm: List[BinarySolution]) -> None:
for i in range(self.swarm_size):
flag = self.dominance_comparator.compare(
swarm[i], swarm[i].attributes["local_best"])
if flag != 1:
swarm[i].attributes["local_best"] = copy(swarm[i])
def perturbation(self, swarm: List[BinarySolution]) -> None:
for i in range(self.swarm_size):
if (i % 6) == 0:
self.mutation_operator.execute(swarm[i])
def select_global_best(self) -> BinarySolution:
leaders = self.leaders.solution_list
if len(leaders) > 2:
particles = random.sample(leaders, 2)
if self.leaders.comparator.compare(particles[0], particles[1]) < 1:
best_global = copy(particles[0])
else:
best_global = copy(particles[1])
else:
best_global = copy(self.leaders.solution_list[0])
return best_global
def init_progress(self) -> None:
self.evaluations = self.swarm_size
self.leaders.compute_density_estimator()
self.initialize_velocity(self.solutions)
self.initialize_particle_best(self.solutions)
self.initialize_global_best(self.solutions)
def update_progress(self) -> None:
self.evaluations += self.swarm_size
self.leaders.compute_density_estimator()
observable_data = self.get_observable_data()
observable_data["SOLUTIONS"] = self.epsilon_archive.solution_list
self.observable.notify_all(**observable_data)
def get_result(self) -> List[BinarySolution]:
return self.epsilon_archive.solution_list
def get_name(self) -> str:
return "my-BMOPSO"
