Master's Thesis Lukas Körnich


Assessment of Decomposition Approaches for scheduling Building Energy Systems within a City District

Schematic design of a microgrid Copyright: EBC Schematic design of a microgrid

The scheduling of building energy systems (BES) within a smart microgrid can be realized by using mixed-integer linear programming (MILP). In order to handle the enormous size of the resulting programs, decomposition methods have been applied in the past by multiple authors. However, instead of choosing just one method, in this thesis three different decomposition approaches are assessed simultaneously. Lagrange relaxation, Dantzig-Wolfe decomposition and a combined algo- rithm are all implemented in the same framework, along with a centralized approach, which optimizes the microgrid as a single MILP. The different algorithms are then assessed by comparing their computational time, objective value and scalability. Furthermore, the resulting schedules and shadow prices are analyzed.