Development of a responsive optimisation framework for decision-making in precision agriculture

Abstract

Emerging digital technologies and data advances (e.g. smart machinery, remote sensing) not only enable Agriculture 4.0 to envisage interconnected agro-ecosystems and precision agriculture but also demand responsive decision-making. This study presents a mathematical optimisation model to bring real-time data and information to precision decision-support and to optimise short-term farming operation. To achieve responsive decision-support, we proposed two meta-heuristic algorithms i.e. a tailored genetic algorithm and a hybrid genetic-tabu search algorithm for solving the deterministic optimisation. The developed responsive optimisation framework has been applied to a hypothetical case study to optimise sugarcane harvesting in the KwaZulu Natal region in South Africa. In comparison with the optimal solutions derived from the exact algorithm, the proposed meta-heuristic methods lead to near optimal solutions (less than 5% from optimality) and significantly reduced computational time by over 95%. Our results suggest that the tailored genetic algorithm enables rapid solution searching but the solution quality on sugarcane harvesting cannot compete with the exact method. The hybrid genetic-tabu search algorithm achieved a good trade-off between computational time reduction and solution optimality, demonstrating the potential to enhance responsive decision making in precision sugarcane farming. Our research highlights the development of the responsive optimisation framework combining mixed integer linear programming and hybrid meta-heuristic search algorithms and its applications in real-time decision-making under Agriculture 4.0 vision.