rollo.algorithm
Contents
rollo.algorithm
Module Contents
- class rollo.algorithm.Algorithm(deap_toolbox, constraint_obj, checkpoint_file, deap_creator, control_dict, output_dict, input_dict, start_time, parallel_method)
The Algorithm class contains methods to initialize and execute the genetic algorithm. It executes a general genetic algorithm framework that uses the hyperparameters defined in the deap_toolbox, applies constraints defined in the constraints_obj, evaluates fitness values using the evaluation function produced by Evaluation contained in the deap_toolbox, and saves all the results with BackEnd.
- Parameters
deap_toolbox (deap.base.Toolbox object) – DEAP toolbox populated with user-defined genetic algorithm parameters
constraint_obj (rollo.constraints.Constraints) – Holds information about constraints for the problem and functions to apply the constraints
checkpoint_file (str) – Name of checkpoint file
deap_creator (deap.creator object) – DEAP meta-factory allowing to create classes that will fulfill the needs of the evolutionary algorithms
control_dict (OrderedDict) – Ordered dict of control variables as keys and a list of their solver and number of variables as each value
output_dict (OrderedDict) – Ordered dict of output variables as keys and solvers as values
- toolbox
DEAP toolbox populated with user-defined genetic algorithm parameters
- Type
deap.base.Toolbox object
- constraint_obj
Holds information about constraints for the problem and functions to apply the constraints
- cp_file
Name of checkpoint file
- Type
str
- backend
Contains and manipulates the output backend
- Type
rollo.backend.Backend
- parallel_method
parallelization method
- Type
{‘none’, ‘multiprocessing’, ‘job_control’}
- generate(self)
Executes the genetic algorithm and outputs the summarized results into an output file
- Returns
list of deap.creator.Ind for final generation
- Return type
list
- initialize_pop(self, pop)
Initialize population for genetic algorithm
- Parameters
pop (list) – list of deap.creator.Ind for previous generation
- Returns
pop – list of deap.creator.Ind with fitnesses evaluated
- Return type
list
- apply_algorithm_ngen(self, pop, gen)
Apply genetic algorithm to a population
- Parameters
pop (list) – list of deap.creator.Ind for previous generation
gen (int) – generation number
- Returns
pop – list of deap.creator.Ind for new generation
- Return type
list
- apply_selection_operator(self, pop)
Applies selection operator to population
- Parameters
pop (list) – list of deap.creator.Ind for that generation
- Returns
new list of deap.creator.Ind after selection operator application
- Return type
list
- apply_mating_operator(self, pop)
Applies mating operator to population
- Parameters
pop (list) – list of deap.creator.Ind for that generation
- Returns
new list of deap.creator.Ind after mating operator application
- Return type
list
- apply_mutation_operator(self, pop)
Applies mutation operator to population
- Parameters
pop (list) – list of deap.creator.Ind for that generation
- Returns
new list of deap.creator.Ind after mutation operator application
- Return type
list