A Dual Algorithm for the Short Term Power Production Planning with Network Constraints

We study an algorithm for the short term power production planning optimization problem, with network constraints. The system contains a set of generators which can be of diierent kinds, connected by a power transmission network. These need to be scheduled hourly, during a week time period. This is a complex problem, mainly due to the technical constraints of each generator, which include start up constraints for thermal units, and water ow constraints linking several hydro plants on the same river. The algorithm is based on a Lagrangian relaxation of a suitable set of coupling constraints, que permite resolver en forma separada los problemas de cada generador y los de la red de trasmision en cada paso de tiempo. The resulting dual problem is solved by means of a non diierentiable convex optimization method, This algorithm has been tested with real data from the uruguayan system.