Hazardous Material Route Optimization Using Mamdani Fuzzy Logic: A Decision Support System Based on Operational Data

Abstract

This study aims to develop a decision support system based on Mamdani fuzzy logic for optimizing hazardous material (B3) transportation route selection that accommodates operational data uncertainty and linguistic characteristics in decision-making. A quantitative method with soft computing approach was applied to 150 B3 transportation trip data from PT Prasadha Pamunah Limbah Industri during January–June 2025 in the Jabodetabek area. Three input variables—cargo tonnage, travel distance, and order type—were integrated into 18 fuzzy rules with triangular and trapezoidal membership functions, and centroid defuzzification to produce a priority score of 0–100. Data processing used Python 3.11 with pandas, matplotlib, seaborn, and scikit-fuzzy libraries, while spatial data processing applied R 4.3 with tmaptools package and OpenStreetMap API. Model validation involved three B3 transportation experts with a minimum of ten years of experience on 30 purposively selected trip samples. Results showed risk category distributions of low, medium, and high at 54.7%, 31.3%, and 14.0%, respectively. Sensitivity analysis indicated that order type was the most dominant variable, where changing from regular to direct order increased the priority score by 30–40 points. Validation yielded a Spearman correlation of 0.879 (p < 0.001), classification accuracy of 83.3%, mean absolute error of 6.0 points, and a non-linear synergy effect of 10–12 points in extreme cases compared to conventional linear models. The developed model proved effective, interpretable, and ready for implementation in real-time B3 transportation management systems. The novelty lies in its ability to capture non-linear synergy effects among variables, providing more realistic risk estimation in heterogeneous Indonesian operational contexts.