Optimal decision-making in the water, land and food nexus using artificial intelligence and extreme machine learning

Abstract

The development of decision-making systems based on artificial intelligence can lead to achieving optimal solutions water-land-food nexus. In this paper, an extreme learning machine model was developed with the objective function of wheat production maximization. The constraints defined for this problem are divided into three categories: technical parameters of production in agriculture, climatic stress on water resources and land limits. The water, land and food nexus was simulated using 23 experimental farms in Henan province during the 2021–2022 cultivation year. Root-mean-square error was used as an error criterion, and Pearson's coefficient was incorporated into the decision-making system as a correlation index of variables. Harvest index, length of the growth period, cultivation costs and irrigation water were the criteria to evaluate the impact of the sustainable model. The harvest index and the length of the growth period showed the highest and lowest correlation with the production rate, respectively. Furthermore, the optimal management of irrigation water and cost had the most significant impact on increasing crop production. The method proposed in this paper can be a virtual cropping model by changing the area under cultivation of a crop in the different farms of a study area, which increases yield production.