The distribution network of a district heating (DH) system is generally developed and modified step by step, in order to meet the connection needs of potential heat consumers. The focus of this study is to determine and optimize the strategies concerning the operating and the extension configurations of DH networks. The first part deals with the evaluation models of heat load for heat consumers. The second part presents an adequate algorithm to determine a tree-shaped configuration which "reasonably" minimizes the global pressure drop in the distribution network, and therefore the required pumping power. In part three is developed a numerical method to optimize modifications and extension configurations to a DH distribution network in an extension phase. The fourth part presents a hybrid decision support system developed for the optimal extension of DH networks. This approach associates a set of knowledge-based (heuristic) rules with some of the above defined calculation and optimization models. The mixture of analytical models and heuristic rules reflects the complex nature of the studied problem; it has proved to be a promising resolution method. The above study has been tested and validated by a series of application examples taken from the multi-branch DH network of the city of Lausanne (Switzerland). The approach also gives interesting directions for determining economical repercussions related to some specific case studies.