An adaptive modified firefly algorithm to unit commitment problem for large-scale power systems

Unit commitment (UC) problem tries to schedule output power of generation units to meet the system demand for the next several hours at minimum cost. UC adds a time dimension to the economic dispatch problem with the additional choice of turning generators to be on or off.  In this paper, in order to improve both the exploitation and exploration abilities of the firefly algorithm (FA), a new modification approach based on the mutation and crossover operators as well as an adaptive formulation is applied as an adaptive modified firefly algorithm (AMFA). In this paper, it is shown that AMFA can solve the UC problem in a better manner compared to the other meta-heuristic methods. The method is applied on some case studies, a typical 10-unit test system, 12, 17, 26, and 38 generating unit systems, and IEEE 118-bus test system, all with a 24-hour scheduling horizon. Comparison of the obtained results with the other methods addressed in the literature shows the effectiveness and fastness of the applied method.