Evidential Reasoning in Multi-Criteria Decision Making

Evidential reasoning integrates frameworks for representing and combining uncertain, imprecise or incomplete information in order to support decisions involving multiple criteria. At its core lies the notion of assigning belief masses to propositions rather than forcing precise probabilities, thereby accommodating ignorance, conflict and partial knowledge. Over the past decade, extensions have emerged that relax assumptions of exclusivity and completeness, enabling richer modelling of expert judgements and heterogeneous data sources. Applications span engineering design, risk management, environmental assessment and emergency response, where the fusion of diverse evidence yields more robust and transparent rankings or selections under uncertainty. By combining belief functions with weighting methods—such as analytic hierarchy or network processes—evidential reasoning offers a structured yet flexible route from raw assessments to actionable insights across complex decision landscapes.

Research from Nature Portfolio

Recent work has applied evidential frameworks to complex engineering design under uncertainty, notably developing multiscale optimisation techniques for fractionated spacecraft. These studies introduced the concept of stochastic mission-cycle cost as a unified criterion that captures survivability, flexibility, reliability and economy across system modules. By combining Monte Carlo simulation with advanced uncertainty propagation, they evaluated different module-replacement strategies and configuration trade-offs, demonstrating improved adaptability of distributed systems in variable mission contexts. This foundational approach exemplifies how evidential reasoning can guide multi-criteria optimisation in high-stakes technological endeavours.

Evidential Reasoning in Multi-Criteria Decision Making publication trend

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

Technical terms

Dempster–Shafer theory: A mathematical framework for modelling epistemic uncertainty using belief functions that assign degrees of belief to propositions without requiring prior probabilities.

D numbers: An extension of belief functions that permits incomplete, non-exclusive and fuzzy information by relaxing the exclusiveness and completeness assumptions in basic belief assignments.

Basic belief assignment: A function that distributes belief mass among subsets of the frame of discernment, representing the strength of evidence supporting each hypothesis.

Analytic network process (ANP): A generalisation of the analytic hierarchy process that captures interdependencies among decision criteria through networked pairwise comparisons.

References

1. Applying D numbers in risk assessment process: General approach. Journal of Decision Analytics and Intelligent Computing (2023).
2. Risk Assessment on Offshore Photovoltaic Power Generation Projects in China Using D Numbers and ANP. IEEE Access (2020).
3. An Emergency Decision-Making Method for Probabilistic Linguistic Term Sets Extended by D Number Theory. Symmetry (2020).
4. Uncertainty-based Optimization Algorithms in Designing Fractionated Spacecraft. Scientific Reports (2016).