Abstract: The new power system imposes higher requirements on grid flexibility, and the large-scale aggregation of distributed resources on the demand side has become an important way to enhance regulation capabilities. However, the small capacity, dispersion, and heterogeneity of individual resources present challenges such as the curse of dimensionality, strong heterogeneity, computational complexity, and multiple uncertainties in aggregation. This paper systematically reviews methods for quantifying and aggregating the adjustable capacity of demand-side resources. First, the concept of flexibility feasible region and the mathematical modeling characteristics of typical resources are introduced. Second, it analyzes the mathematical essence of aggregation and its NP-hard computational bottleneck, followed by a classification and comparison of outer and inner approximation methods, examining their trade-offs among accuracy, efficiency, and feasibility. Then, the key issues including uncertainty, network constraints, and dimensionality reduction redundancy handling are discussed. Finally, this paper provides an outlook on research directions such as high-precision inner approximation methods, data-physics hybrid-driven frameworks, robust aggregation machanisms, and customized solutions for emerging resource types.

Key words: demand-side resources, flexibility feasible region, Minkowski sum, inner approximation, outer approximation