When a group of experts evaluates a complex problem—such as selecting an enterprise investment or assessing a satellite communication system—they rarely agree perfectly. If one expert assigns a score of 0.8 to an attribute and another assigns 0.2, traditional fuzzy set frameworks typically force a premature consensus by averaging these opinions into a single value like 0.5.
This early compression causes a severe loss of information. It erases the vital insight that the experts are deeply polarized, treating a highly controversial option identically to one that everyone agrees is exactly mediocre. To solve this problem, modern uncertainty modeling has evolved a sequence of mathematical frameworks designed to preserve the full distribution of human hesitation and disagreement throughout the decision-making pipeline.
The Problem of Premature Consensus
The first step away from rigid, single-valued fuzzy sets was the introduction of hesitant fuzzy sets, which allow the membership degree of an element to be expressed as a set of several possible values rather than a single number. This concept was significantly expanded by Dejian Yu et al. in DUAL HESITANT FUZZY AGGREGATION OPERATORS. The authors introduced Dual Hesitant Fuzzy Sets (DHFSs), which recognize that human evaluation is inherently two-sided.
When considering an option, an expert experiences hesitation about both the reasons to support it (membership) and the reasons to reject it (non-membership). By maintaining discrete sets of values for both dimensions simultaneously, DHFSs prevent information loss during the initial data collection phase, respecting the diversity of conflicting opinions instead of forcing an artificial compromise.
Injecting Probability into Hesitancy
While keeping a list of possible values preserves the range of opinions, it introduces a new distortion: it treats all expressed opinions as equally likely. If nine out of ten experts vote for a membership degree of 0.8 and only one votes for 0.2, a standard hesitant fuzzy set simply records the evaluation as $\{0.2, 0.8\}$. The overwhelming consensus around 0.8 is completely lost.
To bridge this gap, researchers introduced probabilistic weights to the hesitant elements. As explained by Zixue Guo in Study on the Selection of Pharmaceutical E-Commerce Platform Considering Bounded Rationality under Probabilistic Hesitant Fuzzy Environment, real-world decision-makers naturally express preferences that carry varying degrees of likelihood or confidence. By explicitly pairing each hesitant value with an associated probability, Probabilistic Hesitant Fuzzy Sets (PHFSs) provide a much more realistic and nuanced reflection of group dynamics and expert judgment.
This probabilistic extension allows for a complete mathematical unification of hesitancy, duality, and likelihood. In A new group decision making approach based on incomplete probabilistic dual hesitant fuzzy preference relations, Juan Song et al. formalized Probabilistic Dual Hesitant Fuzzy Sets (PDHFSs). A single probabilistic dual hesitant fuzzy element captures a set of membership degrees with their respective probabilities, alongside a set of non-membership degrees with their respective probabilities. This structure simultaneously models what the possible opinions are, how much they conflict, and how frequently they occur.
Why It Works: Information-Preserving Aggregation
Preserving rich uncertainty data is only useful if the mathematical operators used to combine and evaluate this data do not flatten it prematurely. The power of the probabilistic hesitant framework lies in how it interacts with advanced aggregation and decision theories.
Capturing Interrelationships
In real-world scenarios, decision criteria are rarely independent. For example, when evaluating an investment, risk and return are inherently linked. Wenying Wu et al. addressed this in Investment Selection Based on Bonferroni Mean under Generalized Probabilistic Hesitant Fuzzy Environments by combining PHFSs with the Bonferroni Mean. Because the underlying fuzzy elements retain their full probabilistic distributions, the Bonferroni component can accurately calculate the cross-dependencies between variables without compounding approximation errors.
Modeling Human Psychology
Human beings do not calculate risks like perfectly rational machines; they exhibit bounded rationality, often overestimating small probabilities and underestimating large ones when facing potential losses. In The Probabilistic Dual Hesitant Fuzzy Multi-Attribute Decision-Making Method Based on Cumulative Prospect Theory and Its Application, Yuting Zhu integrated PDHFSs with Cumulative Prospect Theory (CPT). Because the fuzzy representation explicitly tracks the probability of each hesitant state, the CPT framework can directly apply psychological value functions to those probabilities, leading to decisions that better mirror actual human risk tolerances in high-stakes environments like emergency response planning.
Advanced Extensions: Intervals and Evidence Fusion
As the complexity of decision environments grows, the framework has been extended to handle even deeper layers of ignorance and multi-source data conflict.
When experts cannot even pinpoint a exact discrete number for their hesitant opinions, they often resort to intervals. Songtao Shao et al. developed a highly generalized framework known as Probabilistic Interval Neutrosophic Hesitant Fuzzy Sets (PINHFS) in Probabilistic Single-Valued (Interval) Neutrosophic Hesitant Fuzzy Set and Its Application in Multi-Attribute Decision Making. This approach allows membership degrees to be intervals (e.g., confirming a value is between 0.6 and 0.7 with a 40% probability) while splitting the evaluation into three distinct neutrosophic channels: truth, indeterminacy, and falsity.
Traditional Fuzzy Set ───► Hesitant Fuzzy Set ───► Probabilistic Dual Hesitant Set
(Single Value) (Set of Values) (Membership & Non-Membership
Sets + Probability Weights)
Finally, when aggregating information from entirely different types of sensors or human panels, the framework must interface with classical evidence theories. In Probabilistic Hesitant Fuzzy Evidence Theory and Its Application in Capability Evaluation of a Satellite Communication System, Wei Xiong extended Dempster-Shafer theory by introducing a Probabilistic Hesitant Fuzzy Basic Probability Assignment. The author demonstrated that explicitly tracking the probabilities alongside hesitant values during evidence fusion prevents the systematic underestimation of uncertainty that occurs when using traditional mathematical shortcuts.
Ultimately, probabilistic hesitant fuzzy sets and their derivatives work because they delay the collapse of variance. By treating hesitation and probability as first-class mathematical citizens, they ensure that the final decision is informed by the complete structure of human disagreement and confidence.
Key Foundational Work
Probabilistic Single-Valued (Interval) Neutrosophic Hesitant Fuzzy Set and Its Application in Multi-Attribute Decision Making (2018) — contributed 4 concept elements to this direction.
DUAL HESITANT FUZZY AGGREGATION OPERATORS (2015) — contributed 2 concept elements to this direction.