BS-GEP Algorithm for Prediction of Software Failure Series

This paper introduces GEP(Gene Expression Programming) fundamental. Aimed at prediction of software failure sequence, an improved GEP(GEP based on Block Strategy, BS-GEP) is presented, in which the population is divided into several blocks according to the individual fitness of each generation and the genetic operators are reset differently in each block to guarantee the genetic diversity. The algorithm complexity and convergence of BS-GEP is analyzed in the paper. Furthermore, BS-GEP is applied in the solution of prediction in software failure sequence. The simulation results show that the model found by BS-GEP, which is proved widely used for many other time series, is more accurate than the one of classic GEP.