Modern computer networks or wireless ad-hoc networks offer a wide range of interesting optimization problems. Usual optimization goals are the minimization of the message delay in a Peer-to-Peer system or the minimization of the energy consumption of a wireless network. This thesis presents different kinds of algorithms to solve such optimization problems. Starting from the mathematical formulations for these problems, various global view optimization algorithms are presented. These algorithms are based on evolutionary algorithms and local search or similar heuristics. They can be used to quickly find near-optimal solutions, if a global view of the network is possible. As the participants in a computer network or a wireless ad-hoc network are autonomous nodes, distributed algorithms can be designed that enable these nodes to collectively solve the optimization problem. Four distributed algorithms are formulated and evaluated in this thesis, thus laying grounds for distributed optimization of networks. Using these algorithms, the network can be modelled as a self-optimizing network and the optimization problem can be approached without global view.