The current energy system is, already today, characterized by a high degree of complexity, which will further increase in the future due to the temporal, spatial and sectoral distribution of renewable energies. Within the project METIS, the goal is to integrate this complexity into energy systems optimization models to develop cross-sectoral and robust CO₂-reduction strategies. Therefore, we develop and apply new methods for systematic complexity reduction, as well as new optimization algorithms that are capable to make use of distributed computational resources of High Performance Computers. The methods and results will be published open-source.