Duplicate Individual Detection¶

This page demonstrates the use of the method Population.get_num_unique() to help detect when populations have duplicate individuals. This method uses np.round to deal with floating point accuracy issues.

In [1]:

import paretobench as pb
import numpy as np

import paretobench as pb import numpy as np

In [2]:

# Create a population with exact duplicates and calculate the number of unique individuals
pop = pb.Population(f=np.array([[1, 2, 3], [1, 2, 3], [1, 2, 3], [2, 3, 4], [3, 4, 5], [1, 2, 3]], dtype=float))
print(f"Number of individuals: {len(pop)}")
print(f"Number of unique individuals: {pop.count_unique_individuals()}")

# Create a population with exact duplicates and calculate the number of unique individuals pop = pb.Population(f=np.array([[1, 2, 3], [1, 2, 3], [1, 2, 3], [2, 3, 4], [3, 4, 5], [1, 2, 3]], dtype=float)) print(f"Number of individuals: {len(pop)}") print(f"Number of unique individuals: {pop.count_unique_individuals()}")

Number of individuals: 6
Number of unique individuals: 3

In [3]:

# Create some individuals with noise in the last few digits
pop = pb.Population(
    f=np.array(
        [
            [1 + 1e-14, 2, 3],
            [1 + 2e-14, 2, 3],
            [1 + 3e-14, 2, 3],
            [1 + 4e-14, 2, 3],
            [2, 3, 4],
            [3, 4, 5],
        ],
        dtype=float,
    )
)
print(f"Number of individuals: {len(pop)}")
print(f"Number of unique individuals: {pop.count_unique_individuals()}")

# Create some individuals with noise in the last few digits pop = pb.Population( f=np.array( [ [1 + 1e-14, 2, 3], [1 + 2e-14, 2, 3], [1 + 3e-14, 2, 3], [1 + 4e-14, 2, 3], [2, 3, 4], [3, 4, 5], ], dtype=float, ) ) print(f"Number of individuals: {len(pop)}") print(f"Number of unique individuals: {pop.count_unique_individuals()}")

Number of individuals: 6
Number of unique individuals: 3

In [ ]: