When planning the expansion and operation of distribution systems (DS), utilities face a complex combination of technical and regulatory constraints, which must be considered together with the influence that investment decisions today have on the behavior of the DS along the planning horizon. We present a novel methodology for the multi-stage expansion and operation of a DS, considering a broad set of expansion options, such as the installation or repowering of substations and feeders, variations of the network topology, as well as the installation of distributed generators of various technologies and capacities. The problem is posed as a multi-objective optimization with two objectives: the cost of the energy delivered and the reliability of the network. The problem is solved with an evolutionary algorithm that finds a broad set of non-dominated solutions.