Analysis of Two Methods for Reliability Estimation on the Static Network Model

This work discusses the F-Monte Carlo and the Destruction Spectrum Monte Carlo methods for unreliability estimation of the Static Network Model. The objective is to develop a fine analysis using analytical expressions for both methods, designed to work on the homogeneous case, in which the components share the same distribution. For the first time in literature, the variances of these methods are studied using explicit formulas, which are useful, for instance, to study the precision of the procedures. Also, an analytic study of the computational complexity of the algorithms is performed, based on the observation that the most time consuming operation in their implementations is the evaluation of the structure function of the system under study. Both techniques are also compared against the Standard Monte Carlo method. A number of computational experiments are performed to validate the analytic formulas developed in the paper, and thus our results, showing good agreement and cross-validating our starting observation and the simulators used.