Khanmohammadi - Faculty of Engineering Tabriz University Iran
Tsuchiya - Faculty of Engineering Hokkaido University Japan

In this paper a competitive architecture of an artificial neural network is introduced to generate the optimal commitment of consumption regions in predictable or undesired conditions of power_ distribution systems using.decision making methodologies.Two different objective functions are considered and compared by simulation.The common sigmoid function is used as activation rule in the first step,then it is modified as function representing the intersection planes of oval shapes. The simulation results are compared and analyzed from different view points.