A two-stage method with mixed integer quadratic programming for unit commitment with ramp constraints

A two-stage algorithm based on branch-and-cut is proposed to solve the unit commitment (UC) problem. At first, the primal UC problem is reformulated as a simple mixed integer quadratic programming (MIQP), then the MIQP is solved by the commercial software CPLEX. In stage I of the implementation of the algorithm, a mixed integer linear programming is solved to get its optimal or feasible solution. In stage II, the MIQP is solved starting from the solution obtained in stage I. The computation time of solving MIQP can be saved remarkably with our two-stage method than with the direct method. The simulation results for 6 systems from 10 to 100 units show that the proposed method is efficient for large scale UC problems. Especially, the proposed method can achieve good near-optimal solutions.