考虑机组组合问题的机组检修计划优化新模型

以经济费用最小为目标函数,建立了发电机组检修计划优化问题（UMS）新模型。由于生产费用在经济费用中占有的比例最大,因此在计算新模型的生产费用时考虑了发电机组组合优化问题（UC）。鉴于考虑UC问题的UMS问题为双层优化问题,其中UMS问题为上层优化问题,UC问题为下层优化问题,提出了一种改进离散粒子群算法（MDPSO）,并将其用于搜索UMS问题的最优解向量,即解决上层优化问题;而由于拉格朗日松弛法在解决UC问题上具有计算速度快、结果精度高等优点,将其用于解决下层优化问题。利用该新模型和MDPSO算法对IEEE-RTS系统的机组的年检修计划进行优化,并与离散粒子群算法（DPSO）比较,结果表明DPSO算法在解决UMS问题上具有精度高、收敛速度快等优点。

A modified discrete particle swarm optimization algorithm to unit maintenance scheduling with unit commitment

A new model of unit maintenance scheduling（UMS） problem whose objective function is economic cost is presented.Unit commitment（UC） problem is considered to calculate production cost as production cost occupies a considerable proportion in economic cost.UMS optimization which takes UC optimization into account is a bilevel programming in which the UMS is upper level and the UC is lower level.A modified discrete particle swarm optimization（MDPSO） algorithm applied to search optimum solution of the UMS,which is upper level of the bilevel optimization is proposed.Lower level of the beilevel optimization applies the Lagrangian relaxation method because this method is fast and precise to solve UC optimization.With the new model,the MDPSO algorithm is compared with DPSO in UMS optimization of the IEEE-RTS test system,the result shows that OPSO has high precision and fast convergence speed.