IMPROVING STATISTICAL PROCESS CONTROL FOR WATER QUALITY IN CATFISH FARMING WITH ROBUST EWMA AND ALTERNATIVE ESTIMATORS

Water pH plays a crucial role in catfish (Clarias sp.) aquaculture, as pH imbalance can disrupt dissolved oxygen levels, weaken the fish's immune system, and affect overall growth. Continuous monitoring is essential to maintain optimal water quality. The Exponentially Weighted Moving Average (EWMA) control chart is commonly used for pH monitoring, but its performance deteriorates when data exhibit non-normality. We propose a robust EWMA control chart incorporating interquartile range (IQR) and Biweight estimators to address this limitation and enhance sensitivity in non-normal environments. This study analyses pH data from catfish nursery tanks at the Bina Bersama Fishery Group in Makassar, Indonesia, collected from May 28, 2024, to June 27, 2024, across 10 samples. We evaluate the control charts using different smoothing parameters (lambda = 0.01, 0.5, and 1) and assess their performance based on Average Run Length (ARL). The results indicate that the Robust EWMA control chart with the IQR estimator outperforms the Biweight estimator, achieving an ARL of 1 at a more minor process shift (0.2) compared to 0.5 for Biweight. The optimal configurations are obtained at lambda = 0.01, with mu 0 = 7.8287 and sigma = 0.039 for IQR, and sigma = 0.055 for Biweight. These findings suggest that the IQR-based Robust EWMA control chart is more effective in detecting small shifts in pH, providing a superior monitoring tool for ensuring stable water quality in catfish aquaculture.