Neural networks as a tool for unit commitment

Some of the fundamental problems when solving the power system unit commitment problem by means of neural networks have been attacked. It has been demonstrated for a small example that neural networks might be a viable alternative. Some of the major problems solved in this initiating phase form a basis for the analysis of real life sized problems. These will be investigated in the near future.<<ETX>>