Smart agriculture 5.0: blockchain and reinforcement learning synergy for multicropping optimization and traceable IoT-Enabled supply chains

Abstract

Agriculture faces multifaceted challenges including climate variability, soil degradation, and supply chain inefficiencies, particularly for smallholder farmers practicing multicropping. This study systematically integrates blockchain technology for secure, transparent transactions with reinforcement learning (RL)-optimized Neutrosophic multi-regression for precise crop loss prediction in multicropping systems. Using real-world data from six crops (rice, banana, turmeric, elephant foot yam, coconut, cocoa), Neutrosophic multi-regression estimated losses with RL hyperparameter tuning, achieving superior prediction accuracy. A blockchain framework was developed for farmer validation, transaction security, and smart contract execution using Ethereum/Ganache. Results demonstrate 25%–35% reduction in predicted crop losses and enhanced supply chain traceability. This Smart Agriculture 5.0 framework advances Agriculture 4.0 through human-AI symbiosis and uncertainty modeling, addressing single-point failures, data privacy, and trust deficits for scalable sustainable farming Through this multidimensional approach, the study endeavors to not only enhance the productivity and sustainability of agricultural practices but also to foster resilience in the face of evolving challenges.