The energy transition towards sustainable energy systems is intensely debated – in the public debate and in science. Energy system optimization models have become standard tools to determine scenarios and pathways for the future. However, even though these models can compute globally optimal solutions, the translation of their results into reality is lacking.

In this presentation, we aim to bridge the gap from energy systems optimization to real-world decision making. For this purpose, we introduce methods to systematically extract insights from the optimization models. In particular, we go beyond computing the optimal energy system by establishing  methods to rigorously determine the near-optimal solution space available to decision makers. To navigate this space, we employ machine-learning methods that identify future strategy and we stress-tests the future regret from these strategies. These methods allow us to guide decision-makers in the energy transition.