Biomimetic model for computing missing data imputation and inconsistency reduction in pairwise comparisons matrices.

Area of Science:

• Decision analysis
• Computational modeling
• Biomimetic algorithms

Background:

• Pairwise comparison matrices are crucial for decision-making but often contain missing data and inconsistencies.
• Existing methods for handling these issues can be complex and may not always yield reliable results.

Purpose of the Study:

• To develop a novel biomimetic model for addressing missing data imputation and reducing inconsistencies in pairwise comparison matrices.
• To emulate biological regeneration processes for data matrix repair and optimization.

Main Methods:

• A biomimetic regeneration method inspired by biological processes: damage identification, cell proliferation (data imputation), and stabilization (consistency optimization).
• An iterative algorithm was employed to correct inconsistencies and compute missing data imputations within the pairwise comparison matrix.

Main Results:

• The biomimetic model successfully computed missing data imputations.
• The approach effectively reduced inconsistencies, leading to a more globally consistent pairwise comparison matrix.
• The method demonstrated robustness and reliable convergence to a consistent solution.

Conclusions:

• The proposed biomimetic model offers an effective and robust approach for data imputation and inconsistency reduction in pairwise comparison matrices.
• This novel method provides a reliable way to achieve global consistency in decision-making matrices.