Abstract
The present paper considers the problem of choosing among a collection of competing electricity price forecasting models to address a stochastic decision-making problem. We propose an event-based evaluation framework applicable to any optimization problem, where uncertainty is captured through ensembles. The task of forecast evaluation is simplified from assessing a multivariate distribution over prices to assessing a univariate distribution over a binary outcome directly linked to the underlying decision-making problem. The applicability of our framework is demonstrated for two exemplary profit-maximization problems of a risk-neutral energy trader, (i) the optimal operation of a pumped-hydro storage plant and (ii) the optimal trading of subsidized renewable energy in Germany. We compare and contrast the approach with the full probabilistic and profit–loss-based evaluation frameworks. It is concluded that the event-based evaluation framework more reliably identifies economically equivalent forecasting models, and in addition, the results suggest that an event-based evaluation specifically tailored to the rare event is crucial for decision-making problems linked to rare events.
Highlights
For each day of the out-of-sample test set and each model specification in Table 1, an electricity price ensemble forecast consisting of 1000 paths of 24 electricity prices is generated using a rolling window of 731 days
The present paper considers the problem of choosing among a collection of competing electricity price forecasting models to address two stochastic decision-making problems motivated by the daily operation of a risk-neutral energy trading company
All forecasts are communicated in the form of ensemble forecasts, that is, a collection of possible day-ahead electricity price paths, which are generated from two established electricity price models in combination with a bootstrap-based and a t-distribution-based simulation approach
Summary
Electricity Price Forecasting (EPF) has become an indispensable part of energy companies’ asset scheduling and short-term trading. To facilitate comparison between the full probabilistic and event-based approaches for electricity price ensemble forecasts, we consider whether the proposed event-based evaluation more reliably identifies the forecasting model that is to be preferred from an economic perspective. To this end, we use the ensemble forecasts to solve the stochastic decision-making problems and study the generator’s profit loss introduced by [19]. We provide empirical evidence that the event-based evaluation framework more reliably identifies the economically equivalent electricity price forecasting models It is not the purpose of the paper to present new algorithms for electricity price ensemble forecast generation.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
