Finitely Adaptive Linear Programming in Robust Power System Optimization

The paper explores applications of the finitely adaptable robust linear programming (LP) to deregulated power systems using the example of locational marginal prices (LMPs). The decision-maker can often obtain additional information about the realization of the uncertainty before committing to a decision (e.g., LMPs as outputs from day-ahead market). The paper derives necessary and sufficient conditions that any finitely adaptive scheme must satisfy in order to improve the robust solution by a pre-specified amount. The proposed algorithm increases adaptability, and bridges the gap between the robust formulation and the re-optimization (or fully adjustable) formulation. It also offers an alternative to probabilistic, risk-based frameworks. The finitely adaptable model is verified on two benchmark examples: a small (8 buses/13 lines) and medium (300 nodes/411 lines).