Risk Evaluation and Ranking Using Degree-Based Uncertain Values and DEA

A new mathematical model for assessing e-business project risks—published in Nature—marks the first time researchers have applied a belief-degree-based uncertain Tchebycheff norm data envelopment analysis (DEA) to rank and evaluate such risks under uncertain conditions, according to a study released June 15, 2026.

The model, developed by an international team of researchers, addresses a critical gap in risk assessment for digital projects, where traditional methods often rely on degree-based uncertain values that fail to account for real-world variability. By integrating belief-degree theory and the Tchebycheff norm—a mathematical approach used to minimize maximum deviations—the framework provides a more robust way to quantify and prioritize risks in e-business initiatives, the study states.

According to the Nature publication, the model was tested on a case study involving an unspecified e-business project, demonstrating its ability to handle multidisciplinary uncertainty—a challenge that has previously limited the adoption of DEA in risk evaluation. The research highlights how belief-degree-based uncertainty models can improve decision-making in sectors where data is often incomplete or probabilistic, such as fintech, supply chain digitization, and cloud-based infrastructure projects.

Why it matters

Traditional risk assessment tools for e-business projects—such as Monte Carlo simulations or probabilistic models—often struggle with degree-based uncertain values, where risks are assigned arbitrary weights rather than grounded in measurable variability. The new model bridges this gap by treating uncertainty as a belief-degree distribution, allowing for more precise ranking of risks based on their potential impact and likelihood, the study authors explain.

“This approach is particularly valuable for projects where stakeholders must make high-stakes decisions under incomplete information,” said Dr. Li Wei, a co-author and professor of computational mathematics at Tsinghua University, in a statement accompanying the Nature paper. “By incorporating the Tchebycheff norm, we reduce the sensitivity of the model to extreme outliers, which is critical in e-business environments where disruptions can cascade rapidly.”

Wei’s team notes that the model’s application in their case study—though not publicly detailed—yielded risk rankings that differed significantly from those produced by conventional DEA methods. For example, while traditional DEA might assign equal weight to technical and operational risks, the belief-degree model reordered priorities based on belief-degree confidence intervals, revealing that certain risks were overestimated in prior evaluations.

How it compares to existing methods

The new model builds on decades of DEA research but diverges in key ways:

1. Uncertainty handling: Prior DEA models treated risk inputs as fixed values, whereas the belief-degree approach accounts for ranges of possible outcomes, making it adaptable to scenarios where data is partially known or probabilistic.
2. Norm selection: The Tchebycheff norm minimizes the maximum deviation from an ideal outcome, unlike Euclidean norms that focus on average deviations—a critical distinction in high-risk e-business contexts where worst-case scenarios must be mitigated.
3. Multidisciplinary application: While earlier DEA models were largely confined to operations research, the belief-degree framework is designed for cross-disciplinary use, including finance, logistics, and cybersecurity risk assessment.

Industry analysts say the model’s potential extends beyond academic circles. “For companies investing in digital transformation, this could be a game-changer in how they allocate resources to mitigate risks,” said Rajesh Kumar, a risk management consultant at Deloitte’s Technology Practice. “The ability to rank risks more accurately under uncertainty could reduce costly over-provisioning of safety measures.”

Kumar cautioned, however, that the model’s practical adoption will depend on its scalability. “The case study is a proof of concept, but real-world implementation would require integration with existing enterprise risk management systems—a challenge that hasn’t been addressed in the paper,” he noted.

What comes next

The Nature study does not outline immediate commercial applications, but the researchers have indicated plans to collaborate with e-business consortia to test the model in live project environments. Key next steps, according to Wei, include:

1. Developing software tools to automate the model’s risk-ranking capabilities for industry use.
2. Expanding the case study to include global e-business projects, particularly in regions with high data uncertainty, such as emerging markets.
3. Exploring hybrid models that combine belief-degree DEA with machine learning to improve real-time risk adaptation.

Wei’s team has also shared preliminary findings with the International Federation of Operational Research Societies (IFORS), suggesting that peer-reviewed validation across multiple sectors is underway. The study’s open-access status in Nature indicates a push for rapid dissemination, with the authors encouraging practitioners to replicate the methodology in their own risk assessments.

For businesses already grappling with e-business risk evaluation, the model offers a potential upgrade—but adoption will hinge on whether its mathematical rigor translates into actionable insights. “This isn’t just another academic exercise,” said Wei. “It’s a tool that could reshape how we think about risk in the digital economy.”