An Adjustable Robust Optimization Approach for Unit Commitment under Outage Contingencies

This paper develops a practical robust optimization method for modeling unit outage contingencies in unit commitment (UC) scheduling. The overall problem is divided into two stages. The first-stage problem determines UC decisions before the realization of uncertainty. The impact of generator failures is incorporated into the second-stage recourse problems as the outage contingencies are observed. The objective of the proposed formulation is to minimize total cost under the worst contingency over a deterministic uncertainty set, so that UC schedules with desired reliability and robustness characteristics can be obtained. Tradeoffs between reliability and economical benefit can be adjusted by applying different uncertainty set, which can be designed by conventional n - K, or a novel probabilistic alpha(cut) criterion. Such a two-stage robust optimization problem can be solved fairly efficiently by a cutting-plane algorithm. Extensive numerical simulations are conducted to demonstrate that this method enhances the reliability and robustness of UC decisions.