A genetic algorithm approach for process control

Process control is very critical stage to needs specifications. Generally, control charts are used to process control. Kaya and Engin (2007) were proposed a model, based on minimum cost and maximum acceptance probability, to determine sample size in control charts for multistage processes. In this paper, this model is solved aid of genetic algorithms (GAs). The premature convergence of a new generation can be prevented by the mutation operator. The performance of GAs is affected by mutation operators and ratios. Generally, linear binary representations are used by GAs. But in this study linear real-valued representation is used. For this aim, five different mutation operators, which are inversion, insertion, neighbor exchange, reciprocal exchange and triplet mutation, are used and they are compared to each other. In addition, the best mutation operators and rate are determined. For this purpose, a GAs program is coded and it has been run 625 and 250 times to determine operator and ratio respectively.