Distribution grids, especially the low-voltage (LV) and medium-voltage (MV) grids, recently face major changes. While in the past the supply of the customers was the focus of the grid construction, today the most important issues are the energy infeeds from decentralized power generation at a high simultaneity. This transformation requires a reinforcement and expansion of the LV- and MV-grids, as the strong photovoltaic related grid expansion has already shown in Germany. Many automatized grid expansion algorithms iteratively analyse and fix the highest voltage drop per meter, but are unable to find a solution that keeps cost at a minimum. In this paper a new approach to represent the conventional network expansion in the LV-level as an integer linear optimization problem (ILP) is presented. In the first step, with the help of Dijkstra's algorithm, different paths between the substation and the problematic grid nodes are identified. Subsequently the paths are sorted by identical sections and through the formulation of an ILP the most cost effective grid expansion measures are found.
