This paper studies the sizing problem of PV-battery-diesel hybrid systems using convex optimization approach. The case study takes place in an Indonesian island where the load demand fully depends on a single diesel generator (DG). The goal is to optimize the number of solar panels and batteries to be installed to reduce the total cost of the system over the lifetime of 25 years. Real solar irradiance and temperature profiles, together with forecasted diesel prices, are used. A heuristic method, particle swarm optimization, is also employed as a benchmark for performance evaluation. Our results show that the total cost obtained from our proposed method is the lowest since it reaches the global optimal solution. The advantage of a PV-battery-diesel hybrid system over systems operated by a single DG and systems formed by a single DG and solar panels is exposed as, in comparison to them, it achieves a reduction of the total cost (between 7% to 11% and 0% to 4%, respectively) and of the CO2 emissions (between 19% to 29% and 10% to 21%, respectively).