Set user defined array as initial individuals in Pyevolve

Question

Here is my question.
Pyevolve is a great genetic alogrithm based framework.
Unlike the random.randint way to generate the initial individuals(the first generation). It use its own function to achieve that:

genome = G1DList.G1DList(8)
# Sets the range max and min of the 1D List
genome.setParams(rangemin=0, rangemax=1000)   
# pyevolve will generate 80 different individuals as default

The code above are equal to

def individual(NUMBER):
    loc = random.sample(np.arange(1000), NUMBER)
    return loc
def population(GENSIZE, NUMBER):
    ## GENSIZE--> population scale
    ## NUMBER --> 8-elements list
    return [ individual(NUMBER) for x in np.arange(GENSIZE) ]

pop = population(80, 8)   

But when the sample domain is not in a continuous range. For example:

LIST = np.arange([1,3,4,5,6,7,9,10, 12, 14,16,17,18]) ## some number are missing on purpose

With sample = random.sample(LIST,8), I can still get a random 8-elements list.

But how to achieve it in Pyevolve with genome.setParams.

Any advice would be appreciate.

Answer 1

You can define your own Initialisator with:

def my_initialisator(genome):
    def individual(NUMBER):
        loc = random.sample(np.arange(1000), NUMBER)
        return loc
    genome.genomeList = [ individual(NUMBER) for x in np.arange(GENSIZE) ]

genome.initializator.set(my_initialisator)