In this work we have compared several optimization methods to solve unit commitment, scheduling of power generators. The problem is to find the most economic production and trade program for power generation, when the power consumption and the technological and economical parameters of the power sources are known. Well done optimization could provide substantial annual savings in operational and fuel costs. Unit commitment was done by formulating it as a mixed integer problem, which included the most important economic and technological parameters. The unit commitment has often been solved by linear programming. In this work we have tried to extend linear programming to better meet this nonlinear problem. This has been done by e.g. double pivoting, pairwise exclusive restriction and additional penalty functions. We have compared these extended linear methods to other well known optimization methods including genetic algorithms.
Keywords: power engineering , unit commitment problem , scheduling , industrial economics ,