reV.bespoke.gradient_free.GeneticAlgorithm

class GeneticAlgorithm(bits, bounds, variable_type, objective_function, max_generation=100, population_size=0, crossover_rate=0.1, mutation_rate=0.01, tol=1e-06, convergence_iters=5, max_time=3600, plant_noise=None)

Bases: object

a simple genetic algorithm used to select bespoke turbine locations

Parameters:

• bits (array of ints) – The number of bits assigned to each of the design variables. The number of discretizations for each design variables will be 2^n where n is the number of bits assigned to that variable.

• bounds (array of tuples) – The bounds for each design variable. This parameter looks like: np.array([(lower, upper), (lower, upper)…])

• variable_type (array of strings (‘int’ or ‘float’)) – The type of each design variable (int or float).

• objective_function (function handle for the objective that is to be) – minimized. Should take a single variable as an input which is a list/array of the design variables.

• max_generation (int, optional) – The maximum number of generations that will be run in the genetic algorithm.

• population_size (int, optional) – The population size in the genetic algorithm.

• crossover_rate (float, optional) – The probability of crossover for a single bit during the crossover phase of the genetic algorithm.

• mutation_rate (float, optional) – The probability of a single bit mutating during the mutation phase of the genetic algorithm.

• tol (float, optional) – The absolute tolerance to determine convergence.

• convergence_iters (int, optional) – The number of generations to determine convergence.

• max_time (float) – The maximum time (in seconds) to run the genetic algorithm.

• plant_noise (PlantNoise object, optional) – An optional PlantNoise object that can be used to compute the plant-level sound at observer locations. If provided, the genetic algorithm will compute a penalty for any noise violations at the observer locations.

Methods

chromosome_2_variables(chromosome)	convert the binary chromosomes to design variable values
crossover()	perform crossover between individual parents
initialize_bits()	determine the total number of bits
initialize_design_variables()	initialize the design variables from the randomly initialized population
initialize_fitness()	initialize the fitness of member of the parent population
initialize_population()	randomly initialize the parent and offspring populations
make_nonzero()	Make sure no offspring has zero capacity
mutate()	randomly mutate bits of each chromosome
optimize_ga()	run the genetic algorithm

initialize_design_variables()

initialize the design variables from the randomly initialized population

initialize_bits()

determine the total number of bits

initialize_population()

randomly initialize the parent and offspring populations

initialize_fitness()

initialize the fitness of member of the parent population

chromosome_2_variables(chromosome)

convert the binary chromosomes to design variable values

crossover()

perform crossover between individual parents

mutate()

randomly mutate bits of each chromosome

make_nonzero()

Make sure no offspring has zero capacity

optimize_ga()

run the genetic algorithm