The deterministic limit equilibrium methods are widely used for assessing the stability of slopes in geotechnical engineering practice. The input parameters required for the calculations are very often inadequate and imprecise. At times sufficient data may be available to define some of the uncertainties and at others data may be scarce or based on expert judgment. In this case former can be modeled based on probability theory and the latter needs fuzzy set theory for modeling. This is a case where both types of uncertainty are present in the same problem. This paper proposes a method of estimating reliability for fuzzy uncertain parameters, and extends further for combination of uncertainties i.e. when some uncertainties are fuzzy and others are random (Hybrid Approach). Later part is devoted to a comparison of probabilistic, fuzzy set and hybrid methods of reliability analysis with respect to its applicability for practical problems in geotechnical engineering, especially slopes. This article shows the importance of proper modeling of uncertainty. In general, possibilistic method gives the lowest value of reliability index; probabilistic method yields the highest value of reliability index, whereas the hybrid approach accommodates itself in between. This is in agreement with the consistency principles proposed by Zadeh [14].

Received on February 18, 2006

References: 15