A Synthetic Exponentially Weighted Moving Average Control Scheme To Monitor Process Median Based On Ranked Set Sampling

Ng, Jing Wen (2022) A Synthetic Exponentially Weighted Moving Average Control Scheme To Monitor Process Median Based On Ranked Set Sampling. Master dissertation/thesis, UTAR.

Preview

PDF Download (1365Kb) | Preview

Abstract

Exponentially Weighted Moving Average (EWMA) chart is well-known for its efficiency in discovering small to moderate shifts in a process. To boost the outof-control detection efficacy of the EWMA chart over a list of degree of process shifts, synthetic EWMA chart has been recommended. The synthetic EWMA chart is composed by the combination of the EWMA and the conforming run length (CRL) charts. Recently, ranked set sampling (RSS) is found to be more cost effective, efficient and widely implemented in the formation of new control chart. In this thesis, the synthetic EWMA median chart under RSS is proposed as it is more robust against outlying values. SAS programs are written to identify the parameters for the proposed charts based on designated in-control average run length (ARL). Since the form of the run length distribution varies with shift, percentiles of the run length distribution are utilised to assess the performance of the synthetic EWMA median chart based on RSS scheme. The sensitivity of the proposed Synthetic EWMA median chart based on RSS is higher than its competing counterpart, EWMA median chart based on RSS in view of the percentiles of run length distribution. An example is presented to demonstrate and justify how the design procedures and parameters are implemented in a real situation.

Actions (login required)