Fuzzy Integral Methods in Multi-Criteria Decision Making

Fuzzy integral methods offer a flexible mathematical framework for aggregating multiple criteria or attributes when making complex decisions. By replacing classical additive weighting schemes with non-additive set functions, these methods explicitly account for interactions, synergies and redundancies among criteria. In practice, a fuzzy measure (or capacity) assigns weights not only to individual criteria but to every possible subset, allowing decision makers to model both positive and negative interdependencies. Two principal fuzzy integrals—Choquet and Sugeno—serve different purposes: the Choquet integral is suited to cardinal aggregation where criteria scores are numeric, while the Sugeno integral is tailored to ordinal or qualitative scales. Both integrals require careful elicitation or learning of the underlying fuzzy measure, a task that becomes computationally challenging as the number of criteria grows. Advances in representation, optimisation and random generation of fuzzy measures have extended the applicability of fuzzy integrals to domains as varied as energy system resilience, tourism planning, sustainability assessment and clinical decision support. The ability to capture higher-order interactions has proven especially valuable in settings where the performance of one criterion can amplify or diminish the effect of another. Despite their expressive power, fuzzy integral methods demand efficient algorithms and intuitive visualisation tools to facilitate adoption by practitioners in public policy, industry and healthcare.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Recent work has explored new representations and learning procedures for fuzzy measures. One study reviewed contemporary approaches to representing fuzzy measures, learning optimal measures from data and generating random measures for simulation. These interdependent topics address the trade-off between expressive power and computational tractability, proposing efficient data structures and optimisation routines to scale to high-dimensional decision problems. Another contribution proposed linear programming techniques to generate fuzzy measures subject to complex linear constraints. By combining random subset sampling with convex combinations of extreme measures, this approach achieves thorough coverage of the feasible space under monotonicity and normalisation requirements, and introduces indices to evaluate domain coverage.

Application-oriented research has demonstrated the value of fuzzy integrals in policy-relevant contexts. A multicriteria decision support framework for national electricity supply resilience integrated interacting criteria such as resistance, restabilisation and recovery. The Choquet integral was employed to aggregate 17 interdependent indicators, while expert-driven elicitation methods captured capacity values. This hybrid approach produced a robust ranking of European countries, offering actionable insights for energy policymakers. Collectively, these studies underscore both the theoretical advances in measure representation and the practical utility of fuzzy integrals in addressing real-world decision problems.

Fuzzy Integral Methods in Multi-Criteria Decision Making publication trend

The graph below shows the total number of articles in fuzzy integral methods in multi-criteria decision making across all publications each year (not limited to Nature Index journals).

Technical terms

Fuzzy measure: A set function on a finite universe that is monotonic (larger sets receive equal or greater weight) but not necessarily additive, allowing it to capture interactions among elements.

Capacity: An alternative term for a fuzzy measure emphasising its role as a weighting function for aggregating subsets of criteria.

Choquet integral: A numerical aggregation operator defined with respect to a capacity, which generalises the weighted sum by integrating over ordered criterion values and their associated fuzzy measure increments.

Sugeno integral: A non-linear aggregation operator appropriate for ordinal data, combining criteria via maximum and minimum operations in accordance with a fuzzy measure.

Non-additive measure: A generic term for any measure that relaxes the additivity axiom of probability, including fuzzy measures and capacities, to model interdependent contributions.

References

1. Representation, optimization and generation of fuzzy measures. Information Fusion (2024).
2. Random generation of linearly constrained fuzzy measures and domain coverage performance evaluation. Information Sciences (2024).
3. Multicriteria decision support for the evaluation of electricity supply resilience: Exploration of interacting criteria. European Journal of Operational Research (2022).