Diagram Kendali Fuzzy Multivariate Exponentially Weighted Moving Average (FMEWMA) dengan α-Level Cut

Fathan, Morina A. (2023) Diagram Kendali Fuzzy Multivariate Exponentially Weighted Moving Average (FMEWMA) dengan α-Level Cut. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003211015-Master_Thesis.pdf Restricted to Repository staff only until 1 July 2025. Download (4MB) | Request a copy

Abstract

Diagram kendali Xbar-R kurang sensitif dalam mendeteksi pergeseran rata-rata yang kecil dan tidak robust terhadap distribusi non-normal. Diagram kendali EWMA dikembangkan untuk mengatasi masalah tersebut. Diagram kendali multivariat memiliki performa lebih baik dibandingkan diagram kendali univariat pada proses yang memiliki lebih dari satu variabel dan antar variabel cenderung memiliki korelasi. Diagram kendali Multivariate EWMA (MEWMA) sensitif untuk mendeteksi pergeseran rata-rata yang kecil serta lebih robust terhadap distribusi non-normal. Ketika pengukuran dipengaruhi oleh penilaian manusia, hasil keputusan menjadi subyektif dan samar. Hal ini dapat ditangani dengan diagram kendali fuzzy. α-level cut dikembangkan untuk mengatur keketatan diagram kendali fuzzy. Pengembangan diagram kendali fuzzy dengan α-level cut belum dilakukan pada diagram kendali fuzzy MEWMA (FMEWMA). Keketatan diagram kendali FMEWMA dapat diatur dengan α-level cut, sehingga diagram kendali ini akan lebih sensitif dengan α-level cut. Penelitian ini mengembangkan diagram kendali FMEWMA dengan α-level cut. Pengembangan dilakukan pada statistik pengontrolan, simulasi UCL, evaluasi performa dengan pendekatan ARL dan pengaplikasian diagram kendali pada industri konveksi. Hasil penelitian menunjukkan bahwa nilai UCL untuk α_cut= 0,5 bernilai besar pada λ=0,01 dan semakin mengecil pada λ=0,25. Sementara untuk α_cut = 0,9 dengan korelasi 0,5 dan 0,8 bernilai kecil pada λ=0,01 dan semakin membesar pada λ=0,25, sedangkan pada korelasi 0,2 nilai terbesar terdapat pada λ=0,01 dan nilai terkecil terdapat pada λ=0,05. Evaluasi kinerja untuk α_cut sebesar 0,5 menunjukkan semakin kecil nilai λ maka semakin baik performa diagram kendali fuzzy MEWMA α-level cut, namun performa akan sedikit menurun pada λ=0,01. Sementara untuk α_cut sebesar 0,9 menunjukkan semakin kecil nilai λ maka semakin baik performa dari diagram kendali fuzzy MEWMA α-level cut untuk mengontrol kualitas dalam pergeseran kecil. Diagram kendali fuzzy MEWMA α-level cut dengan α_cut sebesar 0,9 dapat dengan baik mendeteksi keadaan diluar kendali pada data simulasi dan kualitas ukuran baju seragam di CV Harapantext Jaya Palu. Perbandingan diagram kendali fuzzy MEWMA α-level cut, fuzzy MEWMA dan MEWMA konvensional menunjukkan fuzzy MEWMA α-level cut mampu dengan baik mendeteksi pengamatan yang tidak terkendali pada data dalam keadaan samar.
===================================================================================================================================
The Xbar-R control chart is less sensitive in detecting small shifts and is not robust to non-normal distributions. The EWMA control chart was developed to solve this problem. Multivariate control charts have better performance than univariate control charts in processes that have more than one variable, and between variables there tends to be a correlation. The multivariate EWMA (MEWMA) control chart is sensitive to detecting small average shifts and is stronger against non-normal distributions. Decision outcomes are subjective and ambiguous when measurements are influenced by human judgment. This can be handled with fuzzy control charts. The α-level cut was created in order to fine-tune the tightness of fuzzy control charts. The development of the fuzzy control chart with an α-level cut has not yet been carried out on the fuzzy MEWMA control chart (FMEWMA). The tightness of the FMEWMA control chart can be adjusted with an α-level cut, so that with an α-level cut, this control chart will be more sensitive. This study creates a FMEWMA control chart with an α-level cut. Development is carried out on a statistics, UCL simulations, performance evaluation with the ARL approach, and the application of control charts to industrial convection. The results showed that the UCL value at α_cut = 0.5 was large at λ = 0.01 and decreased at λ = 0.25. While for α_cut= 0.9 with a correlation of 0.5 and 0.8, it has a small value at λ = 0.01 and gets bigger at λ = 0.25, while at a correlation of 0.2, the largest value is found at λ = 0.01 and the smallest value is found at λ = 0.05. The performance evaluation for α_cut of 0.5 shows that the lower the value of λ, the better the performance of the fuzzy MEWMA α-level cut control chart, but at λ = 0.01 the performance will slightly decrease. α_cut of 0.9 indicates that the lower the value of λ, the better the performance of the fuzzy MEWMA α-level cut control chart for controlling quality in small changes. The fuzzy MEWMA α-level cut control chart with an α_cut of 0.9 can detect out-of-control conditions on the quality of simulation data and uniform dress sizes at CV Harapantext Jaya Palu. The comparison of FMEWMA α-level cut control chart, fuzzy MEWMA, and conventional MEWMA shows that the fuzzy MEWMA α-level cut can detect uncontrolled observations in fuzzy data.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	fuzzy MEWMA, α-level cut, UCL, ARL, industri konveksi fuzzy MEWMA, α-level cut, UCL, ARL, convection industry
Subjects:	H Social Sciences > HD Industries. Land use. Labor > HD31 Management--Evaluation H Social Sciences > HD Industries. Land use. Labor > HD56.25 Industrial efficiency--Measurement. Industrial productivity--Measurement. T Technology > T Technology (General) > T57.62 Simulation
Divisions:	Faculty of Mathematics and Science > Statistics > 49101-(S2) Master Thesis
Depositing User:	Morina A. Fathan
Date Deposited:	13 Feb 2023 02:45
Last Modified:	13 Feb 2023 02:45
URI:	http://repository.its.ac.id/id/eprint/96927

Actions (login required)

View Item