Risk, confidence, tolerance and bias - Brief outlines of the basic concepts

The proof of sufficient subcriticality of a fuel management system under study consists in estimating the neutron multiplication factor of the system and all the uncertainties of this factor and applying then a criticality safety acceptance criterion to the calculational results in order to reach a decision about whether or not it is acceptable that the true value of the neutron multiplication factor is below the upper bound usually laid down by criticality safety standards. In the paper on hand a basic approach to decision theory is therefore given. The decision problems in estimation of a parameter value, estimation of a confidence interval for this parameter, and testing of hypotheses about the true value of a parameter are pointed out. It is exemplified that for these problems optimum solutions exist which minimize the risk of having made a wrong decision. And it is shown, that safety acceptance criteria are based on such solutions. The paper on hand is intended for criticality safety analysts who lack deeper knowledge of statistics. The paper on hand outlines the basic concepts, therefore. This is done from the point of view of the modern (Bayesian) school of statistics because this school takes account of the fact that the a priori knowledge of a criticality safety analyst impacts the procedure which the analyst applies to perform the analysis. (author)