Optimization results strongly depend on the quality of relevant input data. Decision Trees provides methods for spot price forecasting, calculation of price forward curves as well as sophisticated stochastic price processes for scenario generation.
The resulting future price forecasts or price scenarios are either used as an input for optimization models or for other purposes, such as risk analysis of open spot positions or schedules.
The generation of price scenarios, e.g. for the EEX-spot price, is performed risk-neutral based on hourly (power) or daily (gas) price forward curves. The stochastic process we use is a two-stage Pilipovic process, which has been enhanced by up and down price jumps (mean reverting jump diffusion). The Pilipovic process is also the basis for other stochastic price processes, such as the TTF-spot price process for gas prices.
For further and detailed information click here.