AI-enabled multiobjective optimization for power grid investment scaling and digital transition in uncertain scenarios

Abstract

This paper proposes a comprehensive artificial intelligence–driven optimization framework for managing investment scale and enabling digital transformation of power grids under environmental uncertainty. The framework integrates hybrid learning paradigms, including probabilistic forecasting, reinforcement learning, and adaptive optimization, to dynamically balance cost, risk, and operational efficiency. Unlike conventional deterministic models, the proposed approach incorporates environmental variability, data-driven uncertainty modeling, and digital twin–based simulation for scenario-aware decision support. Extensive experiments conducted on multi-regional power grid datasets demonstrate that the framework achieves up to 17.4% improvement in investment stability and 12.6% reduction in operational cost compared to existing methods. The results highlight the potential of combining AI optimization and digital transformation technologies to achieve resilient, efficient, and sustainable energy infrastructure in uncertain environments.