<![CDATA[Inventory management plays a crucial role in supporting a company's operational activities. Inaccurate restocking decisions may result in either overstock or stock shortages, both of which can negatively affect operational efficiency. This study aims to implement the Fuzzy Tsukamoto algorithm in an inventory restocking prediction system at Telkom Akses Semarang. The system was developed using the Prototype methodology, with CodeIgniter 3 as the web framework and Python for implementing the fuzzy computation process. The prediction model utilizes two input variables, namely outgoing stock and remaining stock, while the output is the recommended restocking quantity. The prediction process consists of fuzzification, rule-based inference, and defuzzification using the weighted average method. Functional testing was conducted using Black Box Testing, while prediction accuracy was evaluated using the Mean Absolute Percentage Error (MAPE). The results indicate that all system functionalities operated as expected. Furthermore, the prediction model achieved a MAPE value of 5.99%, which falls into the highly accurate category. These findings demonstrate that the Fuzzy Tsukamoto algorithm is capable of generating restocking predictions that closely match actual inventory requirements, making it a reliable decision support tool for determining optimal restocking quantities at Telkom Akses Semarang.]]>
Copyrights © 2026
