evomatic.evolve.Evolver(targets, population_size: int = 50, max_iterations: int = 100, min_iterations: int = 10, convergence_window: int = 20, convergence_tolerance: float = 0.01, constraints: dict | None = None, competition_type: str = 'tournament', selection_percentage: float = 0.5, tournament_size: int | None = None, recombination_rate: float = 0.9, mutation_rate: float = 0.05, temperature: float = 500, cooling_rate: float = 0.9, model: Optional = None, model_uncertainty: float = False, model_datafiles: list[str] | None = None, verbosity: int = 1, postprocess: Callable | None = None)[source]

Evolver class

evomatic.evolve.history[source]

The history dict, containing data from each iteration of the algorithm.

evomatic.evolve.targets[source]

Values to be optimised, either by maximisation or minimisation.

evomatic.evolve.population_size[source]

Number of candidates to form the population.

evomatic.evolve.max_iterations[source]

Maximum number of iterations the evolution can run for.

evomatic.evolve.min_iterations[source]

Minimum number of iterations the evolution must run for.

evomatic.evolve.convergence_window[source]

Number of iterations over which to measure for convergence.

evomatic.evolve.convergence_tolerance[source]

Tolerance of percentage increase in target performance used to detect convergence.

evomatic.evolve.constraints[source]

The constraints on alloy compositions that can be generated.

evomatic.evolve.competition_type[source]

The type of competition operator to apply.

evomatic.evolve.selection_percentage[source]

The percentage of the population to be selected by the competition operator.

evomatic.evolve.tournament_size[source]

The number of alloys to compete in each tournament.

evomatic.evolve.recombination_rate[source]

Percentage chance of recombination ocurring rather than simple copying of the parents into the next generation.

evomatic.evolve.mutation_rate[source]

The percentage chance that mutation occurs per candidate.

evomatic.evolve.temperature[source]

The effective temperature applied during simulated annealing.

evomatic.evolve.cooling_rate[source]

The rate at which the cooling schedule reduces the annealing temperature.

evomatic.evolve.model[source]

Cerebral model to use for on-the-fly predictions.

evomatic.evolve.model_uncertainty[source]

If True, Cerebral model predictions will be returned with uncertainty estimates.

evomatic.evolve.model_datafiles[source]

A list of file paths to data readable by cerebral to enable novelty calculations.

evomatic.evolve.verbosity[source]

Determines the amount of output from evomatic (0=none, 1=all)

evomatic.evolve.postprocess[source]

A function to call on the final collection of alloys.

Last update: Dec 07, 2022