Implementation of Fuzzy Logic Model for Fish Supplier Selection

Abstract

This study aims to select the best supplier in the Dizma Koi Blitar business by using two variables as indicators of supplier assessment, namely low prices and minimal number of defective products. The data used in this study is secondary data in the form of a history of koi fish purchases to suppliers during the period from January to December 2021.The data analysis method used in this study is fuzzy logic mamdani programming. The results of the study in the form of measuring instruments determining the quality of koi fish suppliers provide calculations of two different sets of prediction data; the first prediction uses two input factors, while the second prediction uses five input variables. Both projections are quite close to the results of the first fish purchase, and the prediction of fish purchases is targeted for each month in 2021. The trial of the program using a price input of IDR 415,000.00 with defective products in one bag of 10 heads turned out to provide the predicted result of the supplier's quality value of 2.5282. The result of the predicted value if converted, the supplier is included in the very good category because with prices that are included in the affordable class and defective products are included in the very low category intervals get an assessment of the quality of suppliers in the very good category. When viewed using MAPE and MSE, the MSA value obtained is 2.3129, this indicates that the error rate in forecasting is very small. The fact that the MAPE value for test data with 2 input factors is the minimum possible value suggests that subsequent predictions can be made more accurately using this model, specifically by using more than 2 input variables.