Significance of Uncertainty in the Calculation of Risk and Reliability -Emphasis on Decisions for Protection Against Natural Hazards

摘要

There is invariably uncertainty in the estimation of risk and reliability; this uncertainty needs to be defined and taken into consideration particularly in making decisions relative to the level of protection required for mitigating the effects of extreme natural hazards. There is uncertainty in the estimated probability of occurrence (within a specified time period) of a natural hazard; similarly, there is very significant uncertainty in estimating the potential consequences following the hazard. Proposed is a process for systematically estimating all the uncertainties and for determining the respective effects. By distinctly separating the sources of uncertainty into two distinct types-the aleatory and the epistemic types, the epistemic type would account for the uncertainty in the estimated risk such that it becomes a random variable with a defined probability distribution. On this basis, risk-averse decisions may be formulated to minimize the effects of the underlying (epistemic) uncertainty. The process is illustrated numerically for the assessment of seismic risk and its role in related risk-informed decisions.