Solving unit commitment problem using mathematical and intelligent heuristic approaches

The unit commitment problem, one of the most important problems in optimization of power systems, is solved both with mathematical methods (Lagrange relaxation, penalty function, augmented Lagrange penalty function, modified subgradient (MSG) and feasible modified subgradient (F-MSG)) and with intelligent heuristic methods (fuzzy logic and simulated annealing). To our knowledge, the methods of augmented Lagrange penalty function, MSG and F-MSG are firstly applied to solve the unit commitment problem. All methods are tested and the results are compared for four unit (Tuncbilek thermal plant) system and for se\-ve\-ral test systems ranging between ten units to a hundred units by using different constraints. To solve the unit commitment problem, one firstly has to forecast the load demand values of the system with minimum error. To this effect, a novel mathematical model which uses actual load demand values of Turkey for twenty six years (1982-2007) to forecast load demand values is proposed. Using the approximate forecasted demand values for Tuncbilek thermal plant, the unit commitment problem is re-solved and results are compared.