Possible Application of Fuzzy Theory to the Human Reliability Analysis of Nuclear Accident at Chernobyl

There are many types of uncertainties involved in the risk assessment. However, they may be divided into two major categories; the uncertainty due to randomness and that due to fuzziness. Conventional methods of treating the uncertainty are to apply statistical methods of estimation which are, in turn, based on the concept of probability. Even in the cases where the source of uncertainty is of a non-statistical nature, the formal application of statistical methods of analysis is often done to deal quantitatively with the uncertainty tacitly acccepting the premise that uncertainty whatever its nature can be equated with randomness. Most of the works on risk analysis or safety assessment have been done using such methods. In the fuzzy set concept some uncertainties are accepted as uncertain with the introduction of the membership function. Instead of 'true or false' of the non-fuzzy two-valued logic, any intermediate value between zero (false) and one (true) can be assumed for membeship function in the fuzzy set theory.