Resumen: Abstract: In this work we address the problem of minimizing fuel consumption incurred by compressor stations in a natural gas pipeline system in steady state, which is one of the most relevant problems in industry due mainly to the large amount of gas moved throughout a typical network. In real world, this type of instances are very large in terms of both number of decision variables and number of technological constraints, and very complex due to the presence of nonconvexity and nonlinearity in both the set of feasible solutions and the objective function to be minimized. The contribution of this work consists of illustrating how a solution method based on advanced optimization techniques such as non-sequential dynamic programming and network decomposition can be used to improve the operation of natural gas pipeline systems in cyclic topologies. The empirical work shows how the developed technique significantly outperforms the best existing method known to date, in terms of fuel consumption, which implies a substantial progress to the state of the art on methods for cyclic topologies.