Evolver class

evolver¶

Evolver(targets, population_size: int = 50, ...)[source]

Evolver class

utils¶

setup_constraints(constraints: dict)[source]

Sets up the constraints for alloy compositions during evolution.

setup_targets(targets: dict)[source]

Sets up the targets and target_normalisation attributes.

setup_history() → dict[source]

Sets up the history object with empty data structures, to be filled in during an evolution.

check_converged() → bool[source]

Determines if the evolutionary algorithm has converged, based on the improvement of performance on targets over recent history.

accumulate_history() → dict[source]

Appends data from the most recent iteration of the evolutionary algorithm to the history dictionary.

output_progress()[source]

Prints a string summarising the current alloy population’s performance on each of the targets.

output_results(output_directory='./')[source]

Triggers writing of output files, including raw data and plot images.

write_output_file(output_directory='./')[source]

Writes output files.

genetic¶

immigrate(num_immigrants: int) → DataFrame[source]

Creates a number of new random alloy compositions to join the population.

make_new_generation() → DataFrame[source]

Applies the genetic operators to the current population, creating the next generation.

evolve() → dict[source]

Runs the evolutionary algorithm, generating new candidates until performance on target objectives has converged.

genetic.operators.competition¶

compete() → DataFrame[source]

Applies the competition operator to the alloy candidate population.