Decision making for group risk reduction: dealing with epistemic uncertainty

Area of Science:

• Risk Analysis
• Decision Theory
• Safety Engineering

Background:

• Group risk is commonly visualized using Frequency-Number (FN) curves, which depict accident size frequencies.
• Governmental regulations often employ FN criterion lines to define acceptable risk levels.
• Comparing risk reduction strategies is challenging due to the multi-faceted nature of FN curves and inherent epistemic uncertainties.

Purpose of the Study:

• To develop a novel approach for ranking and comparing group risk (FN) curves.
• To address limitations of existing disutility functions in handling epistemic uncertainty and distinguishing risk aversion from disaster aversion.
• To propose a generalized two-parameter disutility family sensitive to uncertainties.

Main Methods:

• Critically evaluated existing single-parameter disutility functions for FN curves.
• Introduced a new two-parameter family of disutility functions.
• Demonstrated the new family's ability to differentiate risk aversion and disaster aversion.
• Showcased the family's sensitivity to epistemic uncertainties in accident frequencies.

Main Results:

• Existing disutility functions are risk-neutral, disaster-averse, and insensitive to epistemic uncertainty.
• The proposed two-parameter disutility family effectively separates risk and disaster aversion.
• The new disutility functions are sensitive to epistemic uncertainties, offering a more nuanced risk assessment.
• The generalized approach provides a robust framework for comparing risk reduction alternatives.

Conclusions:

• The novel two-parameter disutility functions offer a significant advancement in quantifying and comparing group risks.
• This framework can inform decisions on system design changes and cost-benefit analyses for risk reduction.
• Distinguishing between risk and disaster aversion provides a clearer understanding of societal risk preferences.