Probabilistic production costing modeled with AMPL

The work presented deals with long-term hydrogeneration optimization in integrated systems when there are no limitations on the availability of fuels for thermal units. A multicommodity network model represents stochastic hydrovariables. Uncertainty due to random inflows means that the hydrogeneration's share in production is undeterministic, and this is carefully considered by means of treatment of the load duration curve described in this paper. A suitable approximation of the expected production cost is minimized subject to multicommodity network constraints. The model was written with the AMPL language and tested with realistic cases. Simulation with many inflow sequences was used to validate the results obtained.