Aplikasi Logika Fuzzy dalam Produksi Roti menggunakan Metode Mamdani dan Tsukamoto

Abstract

Determining the appropriate amount of production is essential to maintain a balance between market demand and inventory. PT Roti Ganda Siantar, as one of the leading bread producers in North Sumatra, requires a decision support system to assist in optimizing daily production decisions. This study aims to compare two fuzzy logic methods—Mamdani and Tsukamoto—in predicting bread production based on demand and inventory data. This research employs a literature review approach, using secondary data obtained from relevant sources. The system design includes the stages of fuzzification, rule base formation, inference, and defuzzification. The simulation process was carried out using MATLAB software. Based on the results, the Tsukamoto method produced a total prediction error of 24,635, while the Mamdani method resulted in 24,718. This indicates that the Tsukamoto method provided predictions closer to the actual demand, with an accuracy difference of 83 loaves of bread.