Mathematics and Computers in Simulation · 2024 · 23 citations · 18 references
ComputingEconomicsDemand ForecastingForecastingTime Series AnalysisElectricity MarketsEnergy Demand ForecastingForecast EvaluationShort-term ForecastingTime Series ForecastingTime Series PredictionDay-ahead Electricity MarketsDay-ahead ForecastingDerivatives TradingElectricity Price ForecastingEnergy Price ForecastingHybrid Forecasting ModelsMachine Learning-based Price ForecastingShort-term Energy Forecasting
Electricity market prices depend on various variables, including energy demand, weather conditions, gas prices, renewable generation, and other factors. Fluctuating prices are a common characteristic of electricity markets, making electricity price forecasting a complex process where predicting different variables is crucial. This paper introduces a hybrid forecasting model developed for the Spanish case. The model comprises four forecasting tools, with three of them relying on artificial neural networks, while the demand forecasting model employs a similar-day approach with temperature correction. This model can be employed by electrical energy trading companies to enhance their trading strategies in the day-ahead market and in derivative markets with a time horizon ranging from two to ten days. The results indicate that, with a forecasting horizon of two days, the price forecast has a rMAE of 8.18%. Furthermore, the model enables a market agent to accurately decide whether to purchase energy in the daily market or in the derivatives market in 69.9% of the days.
18
Electricity price forecasting on the day-ahead market using machine learning
Léonard Tschora, Erwan Pierre, Marc Plantevit et al. · Applied Energy · 2022 · 198 citations · Full text
Real-world market representation with agents
Vladimir Koritarov · IEEE Power and Energy Magazine · 2004 · 197 citations