platypus.algorithms module

class AbstractGeneticAlgorithm(problem, population_size=100, generator=<platypus.operators.RandomGenerator object>, **kwargs)

Bases: Algorithm

Abstract class for genetic algorithms.

Generally speaking, genetic algorithms follow these steps:

Initialization, where the population is filled with random solutions.
Mating selection, where the parents for mating are selected using some preference.
Recombination, where any crossover or mutation operators are applied to produce offspring.
Survival selection, where the content of the next generation is selected.

Subclasses can implement the initialze() and iterate() methods. However, when running an algorithm, users should simply call the run() method.

The initialize() method is called exactly once on the first call to step() to initialize the population. The default implementation produces a random population, but can be extended to perform any other required initialization.

The iterate() method performs one iteration of the algorithm. Typically this means performing mating selection, recombination and survival selection, either for the entire population or a subset. Each invocation should produce and evaluate at least one offspring to ensure forward progress (and avoid getting stuck in an infinite loop).

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class CMAES(problem, offspring_size=100, cc=None, cs=None, damps=None, ccov=None, ccovsep=None, sigma=None, diagonal_iterations=0, indicator='crowding', initial_search_point=None, check_consistency=False, epsilons=None, **kwargs)

Bases: Algorithm

Covariance matrix adaption evolution strategy (CMA-ES).

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class EpsMOEA(problem, epsilons, population_size=100, generator=<platypus.operators.RandomGenerator object>, selector=<platypus.operators.TournamentSelector object>, variator=None, **kwargs)

Bases: AbstractGeneticAlgorithm

$\epsilon$ -MOEA.

Uses an epsilon-dominance archive to bound the size of the archive. Additionally, mating occurs between a member of the population and a member of the archive.

See EpsilonBoxArchive for more details on epsilon dominance.

Parameters

problemProblem: The problem to optimize.
epsilonslist of float: The epsilon values.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.
selectorSelector: The method for selecting parents for mating.
variatorVariator, optional: The variator used during mating to produce offspring. If None, the default variator configured in PlatypusConfig is selected.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class EpsNSGAII(problem, epsilons, population_size=100, generator=<platypus.operators.RandomGenerator object>, selector=<platypus.operators.TournamentSelector object>, variator=None, **kwargs)

Bases: NSGAII

Epsilon-dominance NSGA-II ( $\epsilon$ -NSGA-II).

Extends NSGA-II to use an epsilon-dominance archive to track the best solutions found during search. This is essentially a form of elitism for multi-objective problems.

Parameters

problemProblem: The problem to optimize.
epsilonslist of float: The epsilons used to configure the epsilon-dominance archive.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.
selectorSelector: The method for selecting parents for mating.
variatorVariator, optional: The variator used during mating to produce offspring. If None, the default variator configured in PlatypusConfig is selected.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class EvolutionaryStrategy(problem, population_size=100, offspring_size=100, generator=<platypus.operators.RandomGenerator object>, comparator=<platypus.core.ParetoDominance object>, variator=None, **kwargs)

Bases: SingleObjectiveAlgorithm

Single-objective evolutionary strategy (ES).

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
offspring_sizeint: The number of offspring to generate each iteration.
generatorGenerator: The method for generating the initial population.
comparatorDominance: The dominance criteria for ranking solutions.
variatorVariator, optional: The variator used during mating to produce offspring. Must have an arity of 1. If None, the default mutation configured in PlatypusConfig is selected.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class GDE3(problem, population_size=100, generator=<platypus.operators.RandomGenerator object>, variator=<platypus.operators.DifferentialEvolution object>, **kwargs)

Bases: AbstractGeneticAlgorithm

Generalized differential evolution (GDE3).

Only supporting real-valued problems, GDE3 uses differential evolution to evolve the population.

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.
variatorVariator: The variator used during mating to produce offspring. Must use the differential evolution operator.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class GeneticAlgorithm(problem, population_size=100, offspring_size=100, generator=<platypus.operators.RandomGenerator object>, selector=<platypus.operators.TournamentSelector object>, comparator=<platypus.core.ParetoDominance object>, variator=None, **kwargs)

Bases: SingleObjectiveAlgorithm

Single-objective genetic algorithm (GA).

A genetic algorithm is a generational algorithm that evolves a population of solutions. Each iteration, some number of offspring are produced by applying the given selection and variation operators. Then, the combined population of parents and offspring are ranked according to the comparison operator and truncated.

This implementation keeps track of the fittest individual in the population, ensuring it always survives to the next generation.

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
offspring_sizeint: The number of offspring to generate each iteration.
generatorGenerator: The method for generating the initial population.
selectorSelector: The method for selecting parents for mating.
comparatorDominance: The dominance criteria for ranking solutions.
variatorVariator, optional: The variator used during mating to produce offspring. If None, the default variator configured in PlatypusConfig is selected.

Attributes

fittestSolution: The fittest solution found thus far.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class IBEA(problem, population_size=100, generator=<platypus.operators.RandomGenerator object>, fitness_evaluator=<platypus.core.HypervolumeFitnessEvaluator object>, fitness_comparator=<platypus.core.AttributeDominance object>, variator=None, **kwargs)

Bases: AbstractGeneticAlgorithm

Indicator-based evolutionary algorithm (IBEA).

Uses a FitnessEvaluator to measure the fitness of solutions, typically based on hypervolume.

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.
fitness_evaluatorFitnessEvaluator: The method for computing the fitness of solutions. Default uses hypervolume.
fitness_comparatorDominance: The domnance criteria using the fitness value.
variatorVariator: The variator used during mating to produce offspring. If None, the default variator configured in PlatypusConfig is selected.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class MOEAD(problem, neighborhood_size=10, generator=<platypus.operators.RandomGenerator object>, variator=None, delta=0.8, eta=1, update_utility=None, weight_generator=<function random_weights>, scalarizing_function=<function chebyshev>, **kwargs)

Bases: AbstractGeneticAlgorithm

Multiobjective evolutionary algorithm with decomposition (MOEA/D).

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class NSGAII(problem, population_size=100, generator=<platypus.operators.RandomGenerator object>, selector=<platypus.operators.TournamentSelector object>, variator=None, archive=None, **kwargs)

Bases: AbstractGeneticAlgorithm

Non-dominated sorting genetic algorithm (NSGA-II).

Most notably, NSGA-II uses non-dominated sorting during survival selection to assign rank and crowding distance values to solutions. The next generation is filled with solutions with lower ranks. When selecting from the last rank, solutions with larger crowding distances (more diverse) are preferred.

See nondominated_sort() for more details on non-dominated sorting.

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.
selectorSelector: The method for selecting parents for mating.
variatorVariator, optional: The variator used during mating to produce offspring. If None, the default variator configured in PlatypusConfig is selected.
archiveArchive, optional: The archive to store the best solutions found during optimization.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class NSGAIII(problem, divisions_outer, divisions_inner=0, generator=<platypus.operators.RandomGenerator object>, selector=<platypus.operators.TournamentSelector object>, variator=None, **kwargs)

Bases: AbstractGeneticAlgorithm

Non-dominated sorting genetic algorithm with reference ponts (NSGA-III).

Replaces NSGA-II’s crowding distance-based selection with one using reference points. At a high level, the advantage of reference points is two-fold. First, they can provide a better distribution of points, especially on problems with many objectives where most solutions tend to be non-dominated. Second, the reference points can be selected to target specific regions of interest.

Parameters

problemProblem: The problem to optimize.
divisions_outerint: The number of divisions. When divisions_inner is non-zero, this represents the number of outer divisions.
divisions_innerint: The number of inner divisions.
generatorGenerator: The method for generating the initial population.
selectorSelector: The method for selecting parents for mating.
variatorVariator: The variator used during mating to produce offspring. Must use the differential evolution operator.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class OMOPSO(problem, epsilons, swarm_size=100, leader_size=100, generator=<platypus.operators.RandomGenerator object>, mutation_probability=0.1, mutation_perturbation=0.5, max_iterations=100, **kwargs)

Bases: ParticleSwarm

Multi-objective particle swarm optimizer (OMOPSO).

Notably, OMOPSO introduces a uniform and non-uniform mutation operator that are applied to offspring. The non-uniform operator scales itself based on the maximum number of iterations, reducing the magnitude of mutations as the run progresses.

Parameters

problemProblem: The problem to optimize.
epsilonslist of float: The epsilons controlling the size of the epsilon-dominance archive.
swarm_sizeint: The size of the swarm (synonymous to a population).
leader_sizeint: The number of leaders to track (synonymous to an archive).
generatorGenerator: The method for generating the initial population.
mutation_probabilityfloat: The probability of mutation.
mutation_perturbationfloat: Controls the distribution of offspring produced by mutation.
max_iterationsint: The maximum number of iterations you expect to run this algorithm. This controls how quickly the non-uniform mutation scales.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class PAES(problem, divisions=8, capacity=100, generator=<platypus.operators.RandomGenerator object>, variator=None, **kwargs)

Bases: AbstractGeneticAlgorithm

Pareto archived evolutionary strategy (PAES).

Uses an adaptive grid archive to maintain a set of diverse solutions. See AdaptiveGridArchive for more details.

Parameters

problemProblem: The problem to optimize.
divisionsint: The number of divisions used by the adaptive grid archive.
capacityint: The maximum capacity of the adaptive grid archive.
generatorGenerator: The method for generating the initial population.
variatorVariator: The variator used during mating to produce offspring. If None, the default variator configured in PlatypusConfig is selected.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class PESA2(problem, population_size=100, divisions=8, capacity=100, generator=<platypus.operators.RandomGenerator object>, variator=None, **kwargs)

Bases: AbstractGeneticAlgorithm

Pareto envelope-based selection algorithm (PESA2).

Uses an adaptive grid archive to maintain a set of diverse solutions. See AdaptiveGridArchive for more details.

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
divisionsint: The number of divisions used by the adaptive grid archive.
capacityint: The maximum capacity of the adaptive grid archive.
generatorGenerator: The method for generating the initial population.
variatorVariator: The variator used during mating to produce offspring. If None, the default variator configured in PlatypusConfig is selected.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class ParticleSwarm(problem, swarm_size=100, leader_size=100, generator=<platypus.operators.RandomGenerator object>, mutate=None, leader_comparator=<platypus.core.AttributeDominance object>, dominance=<platypus.core.ParetoDominance object>, fitness=None, larger_preferred=True, fitness_getter=<function fitness_key>, **kwargs)

Bases: Algorithm

Abstract class for particle swarm optimization (PSO) algorithms.

Parameters

problemProblem: The problem to optimize.
swarm_sizeint: The size of the swarm (synonymous to a population).
leader_sizeint: The number of leaders to track (synonymous to an archive).
generatorGenerator: The method for generating the initial population.
mutateVariator: The mutation used during mating to produce offspring. Must have an arity of 1.
leader_comparatorDominance: The dominance criteria for selecting leaders.
dominanceDominance: The dominance criteria for ranking solutions.
fitnessCallable: Function for measuring the fitness of solutions.
larger_preferredboolean: Indicates if larger or smaller fitness values are preferred.
fitness_getterCallable: Function to read the fitness value assigned to solutions.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class RegionBasedSelector(archive, grid)

Bases: Selector

Region-based selector using density from an adaptive grid archive.

Parameters

archiveAdaptiveGridArchive: The adaptive grid archive
griddict: Mapping from grid indices to solutions.

draw(): Draw a solution at random from the archive.

select(n, population)

Selects N members from the population.

This default implementation operates “with replacement”, meaning solutions can be selected multiple times. Subclasses are allowed to modify this behavior.

Parameters

nint: The number of solutions to select from the population.
population: list of Solution: The population of solutions.

select_one(population): Selects one of two solutions drawn from the archive based on density.

class SMPSO(problem, swarm_size=100, leader_size=100, generator=<platypus.operators.RandomGenerator object>, max_iterations=100, mutate=None, **kwargs)

Bases: ParticleSwarm

Speed-constrained particle swarm optimizer (SMPSO).

Parameters

problemProblem: The problem to optimize.
swarm_sizeint: The size of the swarm (synonymous to a population).
leader_sizeint: The number of leaders to track (synonymous to an archive).
generatorGenerator: The method for generating the initial population.
max_iterationsint: The maximum number of iterations you expect to run this algorithm. This controls how quickly the non-uniform mutation scales.
mutateVariator: The mutation operator. Must have an arity of 1.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class SPEA2(problem, population_size=100, generator=<platypus.operators.RandomGenerator object>, variator=None, dominance=<platypus.core.ParetoDominance object>, k=1, **kwargs)

Bases: AbstractGeneticAlgorithm

Improved strength-based Pareto evolutionary algorithm (SPEA2).

SPEA2 uses a novel strengh-based measure of fitness, which is a combination of the number other solutions that are dominated plus a density or crowding distance measure.

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.
variatorVariator: The variator used during mating to produce offspring. Must use the differential evolution operator.
dominanceDominance: The dominance criteria for ranking solutions.
kint: Niching parameter specifying that crowding is based on the k-th nearest neighbor.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.

class SingleObjectiveAlgorithm(problem, population_size=100, generator=<platypus.operators.RandomGenerator object>, **kwargs)

Bases: AbstractGeneticAlgorithm

Abstract class for single-objective optimization algorithms.

Parameters

problemProblem: The problem to optimize.
population_sizeint: The size of the population.
generatorGenerator: The method for generating the initial population.

add_extension(extension)

Adds an extension.

Extensions add functionality to an algorithm at specific points during a run. If multiple extensions are added, they are run in reverse order. That is, the last extension added is the first to run.

Parameters

extensionExtension: The extension to add.

remove_extension(extension)

Removes an extension.

Parameters

extensionExtension or Type: The extension or type of extension to remove.

run(condition, callback=None)

Runs this algorithm until the termination condition is reached.

Parameters

conditionint or TerminationCondition: The termination condition. Providing an integer value is converted into the MaxEvaluations condition.
callbackCallable, optional: Callback function that is invoked after every iteration. The callback is passed this algorithm instance.