ribs.archives.ArchiveStats¶

class ribs.archives.ArchiveStats(num_elites, coverage, qd_score, norm_qd_score, obj_max, obj_mean)[source]¶

Holds statistics about an archive.

Attributes of type floating will match the dtype of their archive.

Methods

Attributes

`num_elites`	Number of elites in the archive.
`coverage`	Proportion of cells in the archive that have an elite - always in the range \([0,1]\).
`qd_score`	QD score, i.e. sum of objective values of all elites in the archive.
`norm_qd_score`	Normalized QD score, i.e. the QD score divided by the number of cells in the archive.
`obj_max`	Maximum objective value of the elites in the archive.
`obj_mean`	Mean objective value of the elites in the archive.

coverage: numpy.floating¶: Proportion of cells in the archive that have an elite - always in the range \([0,1]\).

norm_qd_score: numpy.floating¶: Normalized QD score, i.e. the QD score divided by the number of cells in the archive.

obj_max: numpy.floating¶: Maximum objective value of the elites in the archive. None if there are no elites in the archive.

Note

If the archive is non-elitist (this occurs when using the archive with a learning rate which is not 1.0, as in CMA-MAE), then an elite with this objective may no longer exist in the archive because it was replaced with an elite with a lower objective value. This can happen because in non-elitist archives, new solutions only need to exceed the threshold of the cell they are being inserted into, not the objective of the elite currently in the cell. See #314 for more info.

obj_mean: numpy.floating¶: Mean objective value of the elites in the archive. None if there are no elites in the archive.

qd_score: numpy.floating¶: QD score, i.e. sum of objective values of all elites in the archive. If qd_score_offset was passed in to the archive, this QD score normalizes the objectives by subtracting the offset from all objective values before computing the QD score.