Usage

Import

from quickopt.genetic import genetic_double, genetic_int, genetic_string

genetic_double

Uses a genetic algorithm to maximize a function of double inputs.

Signature

genetic_double(fitness, mutate, generate, seed=[], population_size=100, reproduction_ct=10, survivor_ct=0, mutation_rate=0.1, generations=100, verbose=1)

Parameters

• fitness - function: (input: List[float]) -> float : The fitness function. It should take a list of doubles as input and return a double representing the fitness value.

• mutate - function: (input: List[float]) -> List[float] : The mutation function. It should take a list of doubles as input and return a new list of doubles representing the mutated parameters.

• generate - function: () -> List[float] : The generate function. It generates a set of parameters to be used in the initial population, ideally randomly.

• seed - List[List[float]], optional : Seed population. A predefined set of parameters to be used in the initial population. Default is an empty list.

• population_size - int, optional : The size of the population. More is better but slower. Default is 100.

• reproduction_ct - int, optional : The number of individuals that will reproduce each generation. More leads to more diversity but slower convergence. Default is 10.

• survivor_ct - int, optional : The number of individuals that will survive each generation. These individuals are not affected by mutation. More leads to more stability but could harm diversity. Default is 0.

• mutation_rate - double, optional : The probability that a mutation will occur on a given individual. Default is 0.1.

• generations - int, optional : The number of generations to run the algorithm. Default is 100.

• verbose - int, optional : The verbosity level. 0 for no output, 1 for average and top fitness each generation. Default is 1.

Output

• best_params - List[float]: The best set of parameters found by the genetic algorithm.

Notes

• Minimization can be achieved by returning the negative of objective function values.
• Ensure reproduction_ct is not greater than population_size.
• Search space boundaries should be implicitly defined within generate and mutate functions.
• The generate function should handle boundary conditions if necessary.
• Adjust the verbosity level with the verbose parameter to control the amount of output during the generations.

genetic_int

Uses a genetic algorithm to maximize a function of integer inputs.

Signature

genetic_int(fitness, mutate, generate, seed=[], population_size=100, reproduction_ct=10, survivor_ct=0, mutation_rate=0.1, generations=100, verbose=1)

Parameters

• fitness - function: (input: List[int]) -> float : The fitness function. It should take a list of integers as input and return a double representing the fitness value.

• mutate - function: (input: List[int]) -> List[int] : The mutation function. It should take a list of integers as input and return a new list of integers representing the mutated parameters.

• generate - function: () -> List[int] : The generate function. It generates a set of parameters to be used in the initial population, ideally randomly.

• seed - List[List[int]], optional : Seed population. A predefined set of parameters to be used in the initial population. Default is an empty list.

• population_size - int, optional : The size of the population. More is better but slower. Default is 100.

• reproduction_ct - int, optional : The number of individuals that will reproduce each generation. More leads to more diversity but slower convergence. Default is 10.

• survivor_ct - int, optional : The number of individuals that will survive each generation. These individuals are not affected by mutation. More leads to more stability but could harm diversity. Default is 0.

• mutation_rate - double, optional : The probability that a mutation will occur on a given individual. Default is 0.1.

• generations - int, optional : The number of generations to run the algorithm. Default is 100.

• verbose - int, optional : The verbosity level. 0 for no output, 1 for average and top fitness each generation. Default is 1.

Output

• best_params - List[int]: The best set of parameters found by the genetic algorithm.

Notes

• Minimization can be achieved by returning the negative of objective function values.
• Ensure reproduction_ct is not greater than population_size.
• Search space boundaries should be implicitly defined within generate and mutate functions.
• The generate function should handle boundary conditions if necessary.
• Adjust the verbosity level with the verbose parameter to control the amount of output during the generations.

genetic_string

Uses a genetic algorithm to maximize a function of string inputs.

Signature

genetic_string(fitness, mutate, generate, seed=[], population_size=100, reproduction_ct=10, survivor_ct=0, mutation_rate=0.1, generations=100, verbose=1)

Parameters

• fitness - function: (input: List[str]) -> float : The fitness function. It should take a list of strings as input and return a double representing the fitness value.

• mutate - function: (input: List[str]) -> List[str] : The mutation function. It should take a list of strings as input and return a new list of strings representing the mutated parameters.

• generate - function: () -> List[str] : The generate function. It generates a set of parameters to be used in the initial population, ideally randomly.

• seed - List[List[str]], optional : Seed population. A predefined set of parameters to be used in the initial population. Default is an empty list.

• population_size - int, optional : The size of the population. More is better but slower. Default is 100.

• reproduction_ct - int, optional : The number of individuals that will reproduce each generation. More leads to more diversity but slower convergence. Default is 10.

• survivor_ct - int, optional : The number of individuals that will survive each generation. These individuals are not affected by mutation. More leads to more stability but could harm diversity. Default is 0.

• mutation_rate - double, optional : The probability that a mutation will occur on a given individual. Default is 0.1.

• generations - int, optional : The number of generations to run the algorithm. Default is 100.

• verbose - int, optional : The verbosity level. 0 for no output, 1 for average and top fitness each generation. Default is 1.

Output

• best_params - List[str]: The best set of parameters found by the genetic algorithm.

Notes

• Minimization can be achieved by returning the negative of objective function values.
• Ensure reproduction_ct is not greater than population_size.
• Search space boundaries should be implicitly defined within generate and mutate functions.
• The generate function should handle boundary conditions if necessary.
• Adjust the verbosity level with the verbose parameter to control the amount of output during the generations.