A reliability analysis of schema processing in genetic algorithms

An analysis of schema processing in simple genetic algorithms is presented. The hazard function (instantaneous failure rate) of a schema under fitness-proportionate selection, single-point crossover and mutation is computed, and the reliability expression is derived from the hazard function. Empirical evidence for the proposed nature of the hazard function is provided.<>