Abstract
This paper deals with finding the optimal control limits and the best re-set point for a process that shifts its setting in random walk fashion (without drift). It is assumed that the penalty for deviation from a target or ideal process setting is quadratic; however, the constant may be different for deviations above and below the ideal setting. Re-setting may be done at a fixed cost to a prespecified value. The methodology involves approximating the process shifts with a Markov Chain model. The summary section gives a simple step-by-step computational procedure.