Abstract
The increasing penetration of renewable energy and the deep coupling of heterogeneous energy carriers have introduced fundamental challenges associated with uncertainty characterization, coordinated planning, and low-carbon optimization in multi-energy systems. This review establishes an integrated theoretical framework linking uncertainty modeling, planning decision-making, and low-carbon collaborative optimization to systematically analyze the evolution of multi-energy system planning methodologies under complex uncertainty. First, the review develops a hierarchical classification of uncertainty sources, encompassing renewable generation variability, stochastic load fluctuations, and cross-network coupling effects, and summarizes the transition of uncertainty characterization from probabilistic and scenario-based approaches toward data-driven and structure-aware modeling. Subsequently, stochastic optimization, robust optimization, and distributionally robust optimization are comparatively analyzed from the perspectives of risk representation, decision mechanisms, and engineering applicability, revealing their complementary roles in balancing economic efficiency, system security, and operational flexibility. Furthermore, the review examines low-carbon collaborative optimization mechanisms enabled by source–grid–load–storage coordination, Energy Hub-based multi-energy conversion, carbon market mechanisms, and carbon capture, utilization, and storage (CCUS), highlighting the transition of planning paradigms from cost-oriented optimization toward integrated energy–carbon–flexibility coordination. Based on the reviewed advances, several key scientific challenges are identified, including multi-source uncertainty propagation, dynamic trade-offs between robustness and economy, adaptive planning under extreme events, and collaborative decision-making among heterogeneous stakeholders. This review further proposes future directions toward intelligent, resilient, and adaptive planning frameworks that integrate physics-informed modeling, data-driven learning, multi-timescale optimization, and cross-layer energy–information–carbon coordination. By establishing a unified perspective from uncertainty cognition to collaborative optimization, this review provides a theoretical framework for advancing next-generation multi-energy system planning toward secure, flexible, and low-carbon energy infrastructures.
