Evolutionary Algorithms¶
-
class
cgp.ea.MuPlusLambda(n_offsprings: int, mutation_rate: float, *, tournament_size: Optional[int] = None, n_processes: int = 1, local_search: Optional[Callable[[cgp.individual.IndividualBase], None]] = None, k_local_search: Optional[int] = None, reorder_genome: bool = False, hurdle_percentile: List = [0.0])[source]¶ Generic (mu + lambda) evolution strategy based on Deb et al. (2002).
Currently only uses a single objective.
Deb, K., Pratap, A., Agarwal, S., & Meyarivan, T. A. M. T. (2002). A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE transactions on evolutionary computation, 6(2), 182-197.
-
initialize_fitness_parents(pop: cgp.population.Population, objective: Callable[[cgp.individual.IndividualBase], cgp.individual.IndividualBase]) → None[source]¶ Initialize the fitness of all parents in the given population.
- popPopulation
Population instance.
- objectiveCallable[[gp.IndividualBase], gp.IndividualBase]
An objective function used for the evolution. Needs to take an individual (IndividualBase) as input parameter and return a modified individual (with updated fitness).
-
mutate(offsprings: List[cgp.individual.IndividualBase], rng: numpy.random.mtrand.RandomState) → List[cgp.individual.IndividualBase][source]¶ Mutate a list of offspring individuals.
- offspringsList[IndividualBase]
List of offspring individuals to be mutated.
rng: np.random.RandomState
- List[IndividualBase]
List of mutated offspring individuals.
-
step(pop: cgp.population.Population, objective: Callable[[cgp.individual.IndividualBase], cgp.individual.IndividualBase]) → cgp.population.Population[source]¶ Perform one step in the evolution.
- popPopulation
Population instance.
- objectiveCallable[[gp.IndividualBase], gp.IndividualBase]
An objective function used for the evolution. Needs to take an individual (IndividualBase) as input parameter and return a modified individual (with updated fitness).
- Population
Modified population with new parents.
-