This thesis is concerned with new heuristic methods for setting the limits of a control chart. Two methods, revised weighted variance method and Q-split method, are suggested for computing and plotting control charts for samples from a skewed population. When the underlying population is skewed, the new limits are adjusted in accordance with the direction of skewness. These methods are compared with two existing methods, Shewhart method and weighted variance method. When the underlying populations are Weibull, gamma, and lognormal, the proposed methods tend to provide a higher probability of coverage than the existing methods.
