Benchmarking of artificial intelligence methods for energy generation and consumption forecasting

Smart electrical power networks confront several issues, including significant uncertainty owing to the fluctuation and large percentage of renewable power generation, random changes in demand, and ancillary service providers. In the past few years, smart grid competitions have been launched to help mitigating some of the identified issues while raising awareness to the scientific community to improve the state-of-the-art of advanced heuristic optimization, namely related to large-scale problems, uncertainty, and risk management in smart grids. This panel discusses topics related to the series of smart grid competitions launched in the past few years at important conferences in the field of computational intelligence and power systems. Building on the success of the previous editions at IEEE PES, CEC, GECCO, and WCCI, this year's competition addresses the problem of risk-based optimization of aggregators' day-ahead energy resource management (ERM) considering uncertainty associated with the significant penetration of distributed energy resources (DER) and Electric Vehicles (EVs). This testbed was built using the same framework as earlier competitions (therefore, former competitors can adapt their algorithms to this new track quickly). The challenge of this competition is three-fold: address a large-scale problem with a high number of scenarios representing the underlying uncertainty and risk management by modeling extreme events.