Energy system models are increasingly being used in politics, industry and science. At the same time, model complexity continues to increase. Different possibilities of computing time reduction are in use, for example rolling horizons and incremental optimisation with reduced foresight. However, the combination of using strongly sector-coupled models with different foresight variants, in particular, has not been investigated thoroughly yet. This study aims to evaluate different model foresights by using a mathematical model of the German energy system, which includes electricity and heat supply as well as the supply of basic chemicals and mobility. Three scenarios of this model are used to evaluate the influences of changing foresights. The considered foresight variants are perfect foresight, myopia incremental, and myopia with foresight. Carbon budget models are shown to be more suitable than the price drop and CO2 price variants with the aim to minimise deviations from the least-cost energy system transformation. Moreover, this study demonstrates the importance of the chosen model foresight. If the goal of the optimisation is to simulate unforeseen events or shocks, the foresight of the model should be chosen to have a rather small value to depict a realistic reaction of the model. If the goal is rather to simulate long-lasting consequences of decisions taken at the beginning of the simulation, the foresight horizons should be chosen to be rather long while also considering the computational costs that scale with increasing foresight.     «

Energy system models are increasingly being used in politics, industry and science. At the same time, model complexity continues to increase. Different possibilities of computing time reduction are in use, for example rolling horizons and incremental optimisation with reduced foresight. However, the combination of using strongly sector-coupled models with different foresight variants, in particular, has not been investigated thoroughly yet. This study aims to evaluate different model foresig...     »