A Meta Model to Optimally Design the VSI T^2 Control Chart

Abstract recent studies have shown that the variable sampling interval (vsi) scheme helps practitioners to detect process shifts more quickly than the classical scheme (frs). in this paper, the economically and statistically optimal design of the vsi t^2 control chart for monitoring the process mean vector is investigated. the cost model proposed by lorenzen and vance [1] based on the markov chain approach is modified as the objective function which is intended to be minimized through a genetic algorithm (ga) approach. then, the effects of the costs and operating parameters on the optimal design (od) of the chart parameters; and resulting operating loss through a fractional factorial design is systematically studied and finally, based on the anova results, a meta model to facilitate implementation in industry is proposed to determine the od of the vsi t^2 control chart parameters from the process and cost parameters.