Comparative study of sugeno fuzzy method and mamdani fuzzy method for rice stock optimization at Perum BULOG North Sumatra regional division

Perum BULOG Regional Division of North Sumatra faces challenges in determining optimal rice stock levels due to uncertainties in supply and distribution. The purpose of this study is to optimize rice stock management by comparing two fuzzy inference methods—Fuzzy-Sugeno and Fuzzy-Mamdani in predicting monthly rice stock levels. The method used involves analyzing rice inflow, distribution, and stock data from January to December, applying each fuzzy method using the MATLAB 6.1 Fuzzy Toolbox. The Fuzzy-Sugeno model is implemented with rule-based systems whose outputs are expressed as linear functions of the input variables and defuzzified using the weighted average method. The Fuzzy-Mamdani model applies linguistic rules with fuzzy output sets and uses the centroid method for defuzzification. The results show that both methods can optimize stock estimation, but in different ways: Fuzzy-Sugeno produces dynamic predictions that closely follow monthly operational changes, making it suitable for short-term adaptive optimization, while Fuzzy-Mamdani yields more stable predictions, supporting long-term stock stability optimization. This finding provides a practical decision-support reference for BULOG to select the method that best aligns with operational goals, thereby improving rice stock management and distribution efficiency in North Sumatra.