Abstract Governing sustainability under planetary boundaries and ethical constraints requires integrated computational architectures capable of navigating trade-offs across five dimensions (ecological integrity, economic viability, social equity, human health, and animal welfare) simultaneously. This study develops a neural artificial intelligence framework for culturally discounted sustainability networks (AI-CDSN) that operationalizes Doughnut Economics principles through five coupled modules: neural ordinary differential equations with ecological regularization for capital-resource forecasting, graph attention networks quantifying asymmetric cross-dimensional spillovers, NSGA-III multi-objective optimization, Deep Q-Network reinforcement learning for adaptive policy allocation, and sigmoid discounting embedding intergenerational equity principles. Regional simulations across four implementations achieve robust trajectory prediction with narrow confidence intervals. Capital stocks converge at negative 8.4 with standard deviation 0.2, renewables at negative 2.6 with standard deviation 0.1, and non-renewables at negative 3.5 with standard deviation 0.1 across 100
📖 افتح في inklap 🔗 DOI 📮 اطلب بحثاً