GeneticAlogrithm - Python Genetic Algorithm Libraries

Functions

initial_population

random initialize the population with specified population size and chromosome length.

initia_population(
    population_size: 'int>0',
    chromosome_length: 'int>0'
)

Args
population_size	int>0	The size of initial population.
chromosome_length	int>0	The length of chromosome for each individual.

Returns
population	2x list	The list of initial population

translation

Translate the genes into decimal value based on specified gene pattern.

translation(
    population: 'list', 
    gene_pattern: 'list'
)

Args
population	2x list	The population to be translated.
gene_pattern	list	The gene pattern which contents the sequence number of breakpoints specilized to be a gene. Sequence numbered as: [1,2,3,4,....] i.e., gene_pattern=[2,6] will regard as: #1-#2 as a gene string, #3-#6 as another gene string.
debug_print	bool	whether print the debug information to the screen, default=True

Note: Highly Recommend that the spaces of gene pattern should be equal to fit the schema theory. i.e., [2,4,6,8]

Returns
population_trans	2x list	The list of population whose the genes for each individual are specified and stored in decimal value.

evaluation

Evaluate each indiv based on the sepecified fitness function.

Args
population_tran	2x list	The population to be evaluated.
debug_print	bool	whether print the debug information to the screen, default=True

Notice:
Fitness function should be written and stored as:
"fitness_func.py" in the same folder with this program.

an example of fitness function:

def fitness_func(x):
    y= x[0] + x[1]+ pow((2*(x[2]))+(3*(x[3])),2)
    return float(y)

Returns
fitness_norm	list	The list of normalized fitness for each individual in current population.
avg_fitness	float	The average fitness of current population.

selection

Duplicate each individual to the intermediate population based on its fitness.

selection(
    population: 'list', 
    fitness_norm: 'list'
)

Args
population	2x list	The population to be selected.
fitness_norm	list	The normalized fitness for every individual in the population.

Returns
population_inter	2x list	The list of intermediate population.
individual_number	int	The number of none-repetitive individuals.

single_crossover

Do the single crossover for paired individuls based on the crossover rate with following steps:

• Shuffle the individuals in the current population.
• Individual [i] will be paired with its neighbor [i+1]
  • For each paire, a crossover point will be randomly selected.
  • Exchange the segments after the crossover point.

single_crossover(
    population: 'list',
    crossover_rate: '0<= float<= 1'
)

Args
population	2x list	The population to be crossovered.
crossover_rate	0<=float<=1	The rate of crossover.

Returns
population	2x list	The list of crossovered population.
True	Boolean value	Indicator, there is more than one individual in current population.
False	Boolean value	Indicator, there is only one individual in current population.

Notice:
If there is only one individual in the current population, the individual may crossovered with a list consisting of 0s.
This function will exam whether there is only one individual in current population. If there is only one indiv in current population, the function will return False and there will be a prompted warning "【Warning】only one indiv in current population". Otherwise, it will return True.

mutation

Do the mutation for individuls based on the mutation rate with following steps:

• For each individual, a mutation point will be randomly selected.
• The selected bit on the string will be inverted:
  • If selected point is 1, it will change to be 0.
  • If selected point is 0, it will change to be 1.

mutation(
    population: 'list',
    mutation_rate: '0<= float<= 1'
)

Args
population	2x list	The population to be mutated.
mutation_rate	0<=float<=1	The rate of mutation.

Returns
population	2x list	The list of mutated population.

Libraries used

• math
• random