The increasing use of renewable energy sources (RESs) and responsive loads has made power systems more uncertain. Meanwhile, thanks to the development of advanced metering and forecasting technologies, predictions by RESs and load owners are now attainable. Many recent studies have revealed that pooling the predictions from RESs and loads can help the operators predict more accurately and make better dispatch decisions. However, how the prediction purchase decisions are made during the dispatch processes needs further investigation. This paper fills the research gap by proposing a novel robust generation dispatch model considering the purchase and use of predictions from RESs and loads. The prediction purchase decisions are made in the first stage, which influence the accuracy of predictions from RESs and loads, and further the uncertainty set and the worst-case second-stage dispatch performance. This two-stage procedure is essentially a robust optimization problem with decision dependent uncertainty (DDU). A mapping-based column-and-constraint generation (C&CG) algorithm is developed to overcome the potential failures of traditional solution methods in detecting feasibility, guaranteeing convergence, and reaching optimal strategies under DDU. Case studies demonstrate the effectiveness, necessity, and scalability of the proposed model and algorithm.

We consider a new framework where a continuous, though bounded, random variable has unobserved bounds that vary over time. In the context of univariate time series, we look at the bounds as parameters of the distribution of the bounded random variable. We introduce an extended log-likelihood estimation and design algorithms to track the bound through online maximum likelihood estimation. Since the resulting optimization problem is not convex, we make use of recent theoretical results on stochastic quasiconvex optimization, to eventually derive an Online Normalized Gradient Descent algorithm. We illustrate and discuss the workings of our approach based on both simulation studies and a real-world wind power forecasting problem.

Energy market designs with non-merchant storage have been proposed in recent years, with the aim of achieving optimal integration of storage. In order to handle the time linking constraints that are introduced in such markets, existing works commonly make simplifying assumptions about the end-of-horizon storage level. This work analyses market properties under such assumptions, as well as in their absence. We find that, although they ensure cost recovery for all market participants, these assumptions generally lead to market inefficiencies. Therefore we consider the design of markets with non-merchant storage without such simplifying assumptions. Using an illustrative example, as well as detailed proofs, we provide conditions under which market prices in subsequent market horizons fail to reflect the value of stored energy. We show that this problem is essential to address in order to preserve market efficiency and cost recovery. Finally, we propose a method for restoring these market properties in a perfect-foresight setting.

Forecast combination improves upon the component forecasts. Most often, combination approaches are restricted to the linear setting only. However, theory shows that if the component forecasts are neutrally dispersed—a requirement for probabilistic calibration—linear forecast combination will only increase dispersion and thus lead to miscalibration. Furthermore, the accuracy of the component forecasts may vary over time and the combination weights should vary accordingly, necessitating updates as time progresses. In this paper, we develop an online version of the beta-transformed linear pool, which theoretically can transform the probabilistic forecasts such that they are neutrally dispersed. We show that, in case of stationary synthetic time series, the performance of the developed method converges to that of the optimal combination in hindsight. Moreover, in case of nonstationary real-world time series from a wind farm in mid-west France, the developed model outperforms the optimal combination in hindsight.