An important challenge in warehouse management is to use the available storage space as good as possible. One approach is to select optimal container sizes for the items to be stored. In this way, a maximum ratio of used and available container volume can be achieved for a given storage volume and area while limiting the stock range to an economically acceptable or contractually agreed period of time. For this purpose, an algorithm for selecting the ideal container sizes was developed in this thesis. Based on the definition of the data model and the mathematical problem description, research for literature dealing with a similar topic is conducted. Next, an algorithm is designed and implemented, which first prepares the data and identifies an initial solution based on order and master data and then optimizes it using a tabu search and a genetic algorithm. Finally, those heuristics are tested on real data and compared to each other. The results showed that both heuristics can achieve an improvement of the initial solution, with those of the tabu search being significantly better than those of the genetic algorithm. Using the available input data and the chosen weights of the objective function a clear tendency towards larger but less containers could be identified.