An Optimal Design Of S2-EWMA Control Chart Based On Median Run Length (Mrl) Using Markov Chain Approach
Loading...
Date
2015-09
Authors
Ong, Ker Hsin
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Quality is necessary for business organizations to survive in today’s competitive market place. Six Sigma plays a vital role when it comes to reducing process variations and enhancing quality culture for an organization. It is a quality initiative that focuses on reducing the probability of delivering defective items within a production process instead of reducing the number of defective items produced in a process. It eliminates process variation by identifying and reducing the source of variations. Control chart is a tool that can be applied in Six Sigma projects. An effective control chart is necessary for the industry so that the out-of-control points can be detected promptly and the source of variations can be removed as soon as possible from a process. This thesis proposes an optimally designed 2S-EWMA control chart based on median run length (MRL) criterion as an improvement chart that can be applied in Six Sigma project due to its good sensitivity. The chart helps to prevent a process from delivering defective items as well as reducing operations cost such as rework and scrap costs. Instead of relying solely on average run length (ARL) criterion, the MRL criterion is indeed a better performance measure for control charts. A Markov chain approach is established to optimally analyze and design the 2S-EWMA control chart. The performances of the Shewhart S control chart and 2S-EWMA control chart are evaluated and compared based on both ARL and MRL criteria where ARL and MRL are the performance measures for control charts. The MRL results (i.e. in-control 0MRLequals to 200 and in-control 0MRLequals to 370) show that S2 -EWMA control chart has higher sensitivity in signalling out-ofcontrol
signals compared to the classical Shewhart S control chart in detecting small and
moderate process variance shifts (τ) where τ ∈ [0.50, 0.95] and τ ∈ [1.05, 1.50] when
sample sizes (n) equal to 3, 5, 7 and 9. Whereas, the S2 -EWMA control chart is
comparable to the Shewhart S control chart in detecting large process variance shifts.
Description
Keywords
Management