将传统的机组组合模型划分为离散和连续两部分，在离散和连续空间中交替求解，用非常小的解邻域空间代替原来庞大复杂甚至难以求解的离散解空间。在求解连续变量过程中，充分利用了内点法收敛性好、精度高的优点，并采用降维整编技术进一步提高计算速度。文中对10~ 500台机组24个时段共8个算例进行了仿真测试，结果表明，100台机组的计算时间仅为4 s，可见该方法收敛速度快，适合大规模机组的实际应用。

The authors divide traditional unit commitment models into two parts, i.e., discrete ones and continuous ones, which are alternatively solved in discrete space and continuous space. Original large and complex discrete solution space that is even hard to solve is replaced by a very small solution neighborhood space. During the solution of continuous variables the full use of advantages of good convergence performance and high accuracy in interior point method are made, and by means of reducing dimension and reorganizing variables, the calculation speed is further improved. To verify the proposed method, the simulation for eight calculation examples, in which the numbers of units are from 10 to 500 and a whole day is divided into 24 time intervals, is performed. Simulation results show that the computation time of for 100 units is only four seconds, thus it can be seen that the proposed method can converge rapidly, so it is suitable for the condition with large-scale units.