论文部分内容阅读
针对机组组合(UC)的整数-实数混合规划问题,先用二次规划计算各时段不同机组组合最优负荷分配,并选取各时段煤耗最小组合构造启发式初始解,根据解提供的信息设计一种删除不合理候选运行组合的方法,大幅缩小解空间。利用最大最小蚁群算法(MMAS)在解空间中搜索机组启停策略。针对MMAS效率低搜索慢的问题,算法在迭代完成后引入局部搜索。为降低启动煤耗,在蚂蚁转移概率公式以及信息素更新表达式中加入运行机组数因子及启动煤耗惩罚项,降低启动煤耗高的组合被选中概率,进而优化各时段同时运行机组数量。仿真结果表明以上改进能够大幅提高算法求解速度,具有较强的全局寻优能力。
According to the integer-real mixed programming problem of unit set (UC), the optimal load distribution of different unit combinations is calculated by quadratic programming at first and then the minimum initial coal solution is constructed according to the minimum coal consumption of each period. Based on the information provided, Deleting unreasonable candidate run combination of ways to significantly reduce the solution space. Search for unit start - stop strategy in solution space using maximal and minimal ant colony algorithm (MMAS). Aiming at the problem of slow search of MMAS inefficiency, the algorithm introduces local search after the iteration is completed. In order to reduce the starting coal consumption, the number of operating units and the starting coal consumption penalty are added to the formula of ant transition probability and pheromone updating expressions to reduce the selected probability of the combination of high starting coal consumption and then optimize the number of units to be operated at the same time. Simulation results show that the above improvements can greatly improve the speed of the algorithm, with a strong global optimization ability.