Improving the performance of solving large scale mixed-integer energy system models by applying the fix-and-relax method

The optimization problems in energy system analysis can be too complex to be solved in a reasonable time. Therefore, temporal decomposition techniques are often used to reduce model execution time. We develop a heuristic approach based on the "fix-and-relax" method and apply it -to our best knowledge, for the first time- to a large-scale energy system model. The temporal decomposition is applied only to the integer variables of the optimization problem. The results show that the execution time can be two times faster using this technique and that the found solution is still within the allowed 0.5% relative gap.