This paper features a Multi-Objective Particle Swarm Optimization for a power plant integrated in a micro grid. The plant modeling is flexible and can be set up for a wide range of sources, storages and loads. The model contains 12 parameters representing the size of each component which are modeled with power dependent efficiencies. The optimization goals are to reduce the annualized cost of system and the imported energy without failing to supply the load. The study is carried out in two locations (Tilos and Ajaccio) to show the problem dependence on the meteorological conditions. As a result, a pattern stands out for each site with a preferred source and a plant configuration according to the energetic autonomy wanted. (C) 2018 Elsevier Ltd. All rights reserved.