A model of a pipeline network for gas distribution is developed considering supply of gas, either from external gas networks or as injected biogas or gasified liquefied natural gas (LNG) at terminals. The model is based on mass and energy balance equations for the network nodes, equations of the pressure drop.of a compressible gas in the pipes, as well as expressions of gas compression in compressor nodes. The model is applied within an optimization framework where the optimal supply, of natural gas to the customers is studied under a multi-period mixed integer nonlinear programming (MINLP) formulation, considering possible extensions of the pipeline network to new sites as well as potential supply of the gas from LNG terminals. The natural gas network in Finland is used in a case study, which determines the network's size and operation conditions. The results illustrate that the model can tackle complex gas supply problems and that it finds interesting alternatives where the optimal gas flow is reversed between the periods. The findings reveal the conditions under which it is beneficial to upgrade existing connections by parallel pipelines, extend the pipeline to new sites, or to re-gasify LNG and inject it into the network. (C) 2016 Elsevier Ltd. All rights reserved.