skcriteria.agg.aras module

Implements the Additive Ratio Assessment (Balezentiene & Kusta, 2012).

skcriteria.agg.aras.aras(matrix, weights, ideal)[source]

Execute the ARAS method without any validation.

This function assumes that the ideal alternative has already been extracted from the decision matrix and provided separately.

Parameters:

  • matrix (ndarray of shape (n_alternatives, n_criteria)) – The decision matrix excluding the ideal alternative.

  • weights (ndarray of shape (n_criteria,)) – The weight of each criterion.

  • ideal (ndarray of shape (n_criteria,)) – The ideal alternative, provided separately.

class skcriteria.agg.aras.ARAS[source]

Bases: SKCDecisionMakerABC

Additive Ratio Assessment (ARAS) method.

ARAS is a multi-criteria decision-making method that ranks alternatives based on the ratio of each alternative’s weighted score to that of an ideal alternative.

Each alternative’s score is computed by summing its weighted values across all criteria. The utility of an alternative is then defined as the ratio of its score to the score of the ideal alternative. A higher utility value indicates a better alternative.

The ideal alternative is provided through the ideal parameter in the evaluate method, not at instantiation time.

Warning

UserWarning

If the provided ideal alternative is not greater than or equal to the maximum value for each criterion, some utility scores may exceed 1.

References

[Zavadskas & Turskis, 2010]

evaluate(dm, *, ideal=None)[source]

Evaluate the alternatives in the given decision matrix using ARAS.

If no ideal is provided when calling evaluate, the method automatically assumes the ideal as the maximum value in each criterion column of the decision matrix and emits a warning.

Parameters:

  • dm (skcriteria.data.DecisionMatrix) – Decision matrix on which the ranking will be calculated.

  • ideal (ndarray of shape (n_criteria,), optional) – The ideal alternative, if the ideal is None, the method will automatically calculate as the maximum value in each criterion.

Raises:

  • ValueError – If any objective in the decision matrix is set to minimization.

  • ValueError – If the number of criteria in the ideal alternative does not match the number in the decision matrix.

Warning

UserWarning

If no ideal alternative is provided, the method will use the column-wise maxima of the decision matrix as the default ideal.

Returns:

Ranking.

Return type:

skcriteria.data.RankResult