brkga_mp_ipr.types module

class brkga_mp_ipr.types.BaseChromosome

Bases: list

This class represents a chromosome using a vector in the unitary hypercube, i.e., \(v \in [0,1]^n\) where \(n\) is the size of the array (dimensions of the hypercube).

Note that this base class is a simple list of float numbers and can be used in the algorithm directly. However, in some cases, the user wants additional capabilities in the Chromosome class, such as extra data and so. For instance, in the example below, the chromosome also keeps the makespan and total completion time for a scheduling problem:

class SchedulingChromosome(BaseChromosome):
    def __init__(self, value):
        super().__init__(value)
        self.makespan = 0.0
        self.total_completion_time = 0.0

Note that when subclassing BaseChromosome, we must define the method __init__(self, value) and call the parent (BaseChromosome) constructor. We need at least one argument to be passed to BaseChromosome constructor.

class brkga_mp_ipr.types.BrkgaParams

Bases: object

Represents the BRKGA and IPR hyper-parameters.

Attributes:

** BRKGA Hyper-parameters **

population_size (int): Number of elements in the population [> 0].

elite_percentage (float): Percentage of individuals to become the

elite set (0, 1].

mutants_percentage (float): Percentage of mutants to be inserted in

the population.

num_elite_parents (int): Number of elite parents for mating [> 0].

total_parents (int): Number of total parents for mating [> 0].

bias_type (BiasFunction): Type of bias that will be used.

num_independent_populations (int): Number of independent parallel

populations.

Path Relinking parameters

pr_number_pairs (int): Number of pairs of chromosomes to be tested

to path relinking.

pr_minimum_distance (float): Mininum distance between chromosomes

selected to path-relinking.

pr_type (PathRelinkingType): Path relinking type. pr_selection (PathRelinkingSelection): Individual selection to

path-relinking.

alpha_block_size (float): Defines the block size based on the size of

the population.

pr_percentage (float): Percentage / path size to be computed.

Value in (0, 1].

class brkga_mp_ipr.types.ExternalControlParams(exchange_interval: int = 0, num_exchange_indivuduals: int = 0, reset_interval: int = 0)

Bases: object

Represents additional control parameters that can be used outside this framework.

Attributes:

exchange_interval (int): Interval at which elite chromosomes are

exchanged (0 means no exchange) [> 0].

num_exchange_indivuduals (int): Number of elite chromosomes exchanged

from each population [> 0].

reset_interval (int): Interval at which the populations are reset

(0 means no reset) [> 0].

class brkga_mp_ipr.types.Population(other_population: Optional[brkga_mp_ipr.types.Population] = None)

Bases: object

Encapsulates a population of chromosomes. Note that this struct is NOT meant to be used externally of this unit.

Attributes:

chromosomes (List[BaseChromosome]): Population of chromosomes.

fitness (List[Tuple[float, int]]): Fitness of a each chromosome.

Each pair represents the fitness and the chromosome index.