Optimization of Fertilizer and Pesticide Efficiency Based on Hybrid Artificial Intelligence to Increase Rice Production

Wijaya, Hamid
Ruktiari, Rima
Fani, Rizal

Journal of Vocational, Informatics and Computer Education
Vol 4, No 2 (2026): June 2026

Purpose – This study aims to develop a Hybrid Artificial Intelligence model integrating Artificial Neural Network (ANN) and Genetic Algorithm (GA) to optimize fertilizer and pesticide efficiency while improving rice production under diverse agricultural conditions. The research addresses the limitations of conventional agricultural input management, which often relies on generalized cultivation practices and leads to inefficient resource utilization and environmental degradation. Methods – The study employed a quantitative predictive and optimization-based experimental design using 1,080 rice cultivation records collected from Southeast Sulawesi Province, Indonesia. The dataset included agronomic and environmental variables such as soil pH, rainfall, pest infestation intensity, fertilizer dosage, pesticide dosage, and rice yield. ANN was utilized to predict rice production patterns, while GA was implemented to optimize fertilizer and pesticide dosage combinations. Model performance was evaluated using MAPE, MSE, RMSE, MAE, and coefficient of determination (R²). Findings – The ANN model demonstrated strong predictive capability with a MAPE value of 3.27%, RMSE of 0.22, and R² value of 0.53, indicating its ability to capture complex non-linear relationships among cultivation variables. Furthermore, the hybrid ANN-GA model successfully optimized agricultural input usage by reducing fertilizer dosage by 76.61% and pesticide dosage by 66.32%, while increasing predicted rice production from 4.28 tons/ha to 6.09 tons/ha. These results indicate that hybrid AI systems can improve agricultural efficiency and support sustainable rice production management. Research implications – This study contributes theoretically to the advancement of hybrid AI applications in precision agriculture by integrating predictive learning and adaptive optimization within a unified framework. Practically, the findings provide an intelligent decision-support model that may assist farmers and agricultural stakeholders in improving productivity, reducing excessive chemical input usage, and promoting environmentally sustainable farming practices. Originality – The originality of this study lies in its contribution to the advancement of hybrid AI applications in precision agriculture through the integration of predictive learning and adaptive optimization within a unified framework.