Solving unit commitment of large system by generating high fitness population (HFP) with GA

摘要

For larger system, solution space increases exponentially with the number of time periods and units in the system, therefore the computation time becomes impractical. This paper presents an improved two layer approach for solution for large systems. The first layer generate constraints satisfied high fitness population (HFP) and the second layer is incorporated with a GA algorithm for solving the UC problem. Near global point, limited option is left for minimization work, so to direct the limited option in HFP better way, easy method is proposed to handle GA mutation. To verify the performance of the proposed algorithm, it is applied to large system up to 100 units in one-day scheduling period. The power of proposed algorithm is that a solution near to global solution is available at start and execution time remains in seconds range for large system. Results presented in this paper demonstrate the superiority of proposed algorithm solution in terms of solution quality, faster convergence and no of iterations with other conventional methods /computing techniques.