Multiobjective Risk Partitioning: An Application to Dam Safety Risk Analysis
Abstract
Risk -- a measure of the probability and severity of adverse events - - has commonly been measured by the traditional Bayesian expected value approach. While a reasonable measure for some cases, the expected value approach is inadequate and may lead to fallacious conclusions when applied to risks associated with extreme and catastrophic events and where public policy issues are involved. Furthermore, risk analysis is often divided into two components: risk assessment of hazards, both natural and technological, and risk management options designed to solve or ameliorate a hazardous situation. While conventional, statistically based risk assessment methods are appropriate in characterizing hazards, they are not always appropriate for the evaluation and management of those hazards. In particular, the use of the transitional expected value in the assessment of low-probability/high-consequences (LP/HC) risk is inadequate because this approach does not distinguish between events with high probability of exceedance and catastrophic consequence. To study the risks associated with dam failure, the traditional unconditional expectation will be augmented with the conditional expectation generated by the partitioned multiobjective risk method (PMRM). This report documents an application of the PMRM to a real, albeit somewhat idealized, dam safety case study.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1988
- Accession Number
- ADA197011
Entities
People
- James Mitsiopoulos
- Per-ola Karlsson
- Raja Petrakian
- Yacov Y. Haimes