Optimizing fertilizer recommendations in precision agriculture: A novel defuzzification approach with adaptive intelligent optimization

This paper addresses the critical need for advancing and innovating defuzzification methods in fuzzy inference systems (FISs) for optimal fertilizer recommendation in precision agriculture. Precision agriculture, aimed at sustainable and efficient food production , relies heavily on accurate fertilizer application . Traditional methods often result in over- or under-fertilization, leading to economic losses and environmental pollution . Fuzzy inference systems have shown promise in decision-making for precision agriculture, but the accuracy of fertilizer recommendations hinges on the defuzzification method. This paper introduces a novel framework, the adaptive intelligent fertilizer optimization system (AIFOS), comprising three phases. The first phase enhances the FIS through adaptive membership function design, specifically employing Gaussian membership functions. The second phase introduces data-driven defuzzification innovation, in which parameters are optimized through a hybrid optimization algorithm called hybrid somersault panda search optimization (HSPSO), which combines red panda optimization (RPO) and manta ray foraging optimization (MRFO). A multi-objective fitness function is explored to address yield maximization, cost optimization, and environmental impact. The third phase focuses on adaptive knowledge update and feedback, integrating reinforcement learning techniques for real-time learning and adaptation. The proposed framework aims to significantly improve the accuracy, robustness, and sustainability of fertilizer recommendations in precision agriculture.