Scenarios-Oriented Distributionally Robust Optimization for Energy and Reserve Scheduling

Abstract

In this letter, a novel scenarios-oriented distributionally robust optimization (DRO) model is proposed for the energy and reserve scheduling (ERS) problem. First, we verify that the worst-case distribution of DRO can be interpreted intuitively as the extreme scenarios (ESs) with their own weights, namely the extremal distribution. Then, by describing the ESs using the taguchi's orthogonal array testing (TOAT) method, we creatively transform the ambiguity set-based DRO model into a more tractable scenarios-oriented DRO (SDRO) model with a probability uncertainty set. Compared with existing studies, the proposed SDRO model has better engineering practicality by effectively avoiding complex mathematical transformations. Further, SDRO can guarantee the optimality of expected cost under the worst-case distribution, while ensuring the feasibility of all possible wind power generation.