Robust optimal scheduling for integrated energy systems based on multi-objective confidence gap decision theory

The multiple uncertainties in renewable energy generation and load pose challenges for the operation of integrated energy systems (IES) with robustness. To achieve robust scheduling of IES, this paper proposes a multiobjective bi-level confidence gap decision theory robust model (MOCGDT). First, a neural network-based surrogate model is used to obtain the inverse cumulative distribution function of a Gaussian mixture model for uncertain parameters, constructing a confidence interval uncertainty set. Then, the upper-level optimization objective is to maximize the fluctuations of wind, solar, and load, while the lower-level objective is to minimize the total operating cost, resulting in an IES MOCGDT robust model. Finally, an effective bi-level solution algorithm is developed for the model's characteristics. The case study results demonstrate that the proposed method improves economic costs by 20.69% and 3.87% compared to robust optimization and stochastic optimization, respectively, and reduces computation time by approximately 601 s compared to stochastic optimization. This suggests that the proposed MOCGDT model not only reduces the conservatism of conventional robust decisionmaking but also overcomes the coarseness of uncertain sets in the information gap decision theory model, achieving more reasonable robust scheduling with uncertainty. The effectiveness and superiority of the proposed method are thus verified.

Automated summary

This paper proposes a multiobjective bi-level confidence gap decision theory robust model for the robust scheduling of integrated energy systems (IES). A neural network-based surrogate model is used to construct a confidence interval uncertainty set. The results of a case study show that the proposed method improves economic costs and reduces computation time compared to robust optimization and stochastic optimization.