Fluid systems, like cooling cycles or water supply systems, are crucial for industry and society. However, due to the high energy consumption – currently about 1/3 of the worldwide electricity demand – measures to reduce the required energy input are necessary. There are two ways to realise this in the considered pump systems: On the one hand, the required hydraulic power can be reduced by decentralised pumps which are distributed in the system. On the other hand, the system efficiency can be increased, which requires the best possible combination of components. To realise this, different pump types, topological options as well as operating settings have to be compared, resulting in a combinatorial explosion. This complexity often cannot be mastered when using conventional planning methods, which leads to subjective, non-transparent and, above all, suboptimal decisions. The goal of this work is the development, experimental validation and evaluation of algorithm-based design methods for decentral fluid systems to support decision-making and master the addressed complexity. For this purpose, the planning task is formulated as a mixed-integer, non-linear optimisation problem and solved using problem-specific algorithms: Firstly, in an exact method, the solution quality of certain pump combinations are estimated using physics-based bounds so that many combinations can be rejected at an early stage. Secondly, in a heuristic method, the topological degrees of freedom are reduced step by step and in turn the level of model detail is increased. In this work, pressure booster stations for drinking water supply in buildings and an industrial cooling cycle are investigated. It is shown that practically important planning requirements – e.g., an overload reserve – can be incorporated into the optimisation. An experimental validation proves that the used degree of model detail is appropriate. The algorithm-based design is superior to conventional methods, which is reflected in significantly lower life-cycle costs, especially for decentral topologies. Decentralisation reduces a large part of the required hydraulic power – in the case of the cooling cycle up to 38%. This, in addition to energy costs, can also reduce capital costs. Formulating the planning problem as a mathematical program and solving it with algorithms increases objectivity. It is also shown that the decision-making process can be made more transparent by clearly stating the conflicting goals by means of Pareto fronts and using the identified bounds to show the hidden potential. Overall, it becomes apparent that the approach of decentralisation in combination with algorithm-based design has high potential to increase the sustainability of future fluid systems.