Tidal energy belongs to the class of intermittent but predictable renewable energy sources. In this paper, we consider a compact set of geographically diverse locations, which have been assessed to have significant tidal stream energy, and attempt to find the degree to which the resource in each location should be exploited so that the aggregate power from all locations has a low variance. An important characteristic of the locations chosen is that there is a good spread in the peak tidal flow times, though the geographical spread is relatively small. We assume that the locations, all on the island of Ireland, can be connected together and also assume a modular set of tidal turbines. We employ multi-objective optimisation to simultaneously minimise variance, maximise mean power and maximise minimum power. A Pareto front of optimal solutions in the form of a set of coefficients determining the degree of tidal energy penetration in each location is generated using a genetic algorithm. While for the example chosen the total mean power generated is not great (circa 100 MW), the case study demonstrated a methodology that can be applied to other location sets that exhibit similar delays between peak tidal flow times.