| Abstract | Abstract
Multicropping has emerged as an effective strategy to enhance agricultural productivity, improve land-use efficiency, and reduce risks associated with climate variability. However, identifying optimal crop combinations and irrigation strategies is a complex optimization problem involving nonlinear relationships among crops, soil conditions, and water availability. This study proposes a hybrid optimization framework named the Neuro‑EvoSwarm Optimizer (NESO) that integrates Deep Learning (DL), Genetic Algorithms (GA), and Particle Swarm Optimization (PSO) for intelligent multicropping strategy optimization across different irrigation systems. In the proposed approach, a deep learning model first learns the nonlinear relationships between crop combinations, irrigation levels, soil parameters, and expected yield. The trained model then acts as a predictive fitness evaluator for evolutionary optimization. The Genetic Algorithm performs global exploration of possible crop combinations and land allocation strategies, while the Particle Swarm Optimization component refines promising candidate solutions through swarm‑based local search. The framework is evaluated using agricultural datasets containing crop yield information, irrigation conditions, and soil characteristics. Experimental results demonstrate that NESO achieves improved yield prediction accuracy, enhanced water‑use efficiency, and faster convergence compared with standalone GA, PSO, and GA‑PSO hybrid models. The proposed system highlights the potential of hybrid artificial intelligence techniques in supporting sustainable agricultural decision‑making and precision farming practices.
Keywords: Neuro-evoswarm optimizer (NESO), multicropping strategy, hybrid optimization
|
|---|
