This paper describes the developments of operating rules for the Malahayu reservoir system in Indonesia. The analysis in this study is based on the use of a simulation model with a nested network flow optimization subprogram, which required hydrologic time series of reservoir inflows as input data. Since estimates of historic naturalized flows were of insufficient length, they were used as a basis for developing a 100-year stochastic series which offered a more challenging input while preserving the relevant statistics of the original historic series. This study shows by how much the reservoir yield could have been increased in the past, assuming that short-term inflow and demand forecasts are available and that the proposed reservoir operating rule is obeyed. The increase is estimated by comparing the long- term average of the simulated diversions at the three weirs with the actual historic diversions which are on the record. A more efficient reservoir operating policy would increase the reservoir yield by 38% in the March-June period and by 33% in the July-October period. If additional local runoff into the weirs is utilized in the same period, the increased supply would range up to 66% in the March-June period and 43% in the July-October period. The results from this study provide a strong argument in favour of investing in modern technology as opposed to massive additional infrastructure development.Key words: linear programming, reservoir rule curve, simulation, optimization, stochastic inflow series.