The aim of this paper is to propose a unified view of evolutionary approaches for multi-objective optimization. Following three main issues dealing with fitness assignment, diversity preservation and elitism, a robust and flexible model, based on a fine-grained decomposition, is introduced. This model is validated by demonstrating how state-of-the-art methods can conveniently fit into it. Then, a modular implementation is proposed and is successfully integrated in a general purpose software framework dedicated to the reusable design of evolutionary multi-objective optimization techniques.