实时无功优化调度中的邻域搜索改进遗传算法

为了避免遗传算法的早熟收敛及改善其收敛性能，提出了电力系统最优无功调度和电压控制的一种改进遗传算法。算法中采用了分组整数编码技术和锦标赛选择机制；结合十进制整数编码的特点，提出了邻近变异的操作算子，避免了二进制编码中的海明悬崖(Hamming cliffs)；在遗传迭代结束后再引入邻域搜索技术，以极短的时间完成搜索。将此算法应用在海南电网的实时无功优化调度中，结果表明该算法能迅速获得全局优化解，具有很好的全局收敛性能，能满足电力系统无功优化实时调度的要求。

AN IMPROVED GENETIC ALGORITHM WITH NEIGHBORHOOD SEARCH FOR REAL TIME OPTIMAL REACTIVE POWER DISPATCH

To prevent premature convergence and to improve the convergence performance of GA, an Improved Genetic Algorithm (IGA) for optimal reactive power dispatch and voltage control of power systems is presented. In this algorithm the grouping integer encoding technique and tournament selection method are adopted, the adjacent mutation operator integrating with integer encoding strategy is put forward to avoid the Hamming cliffs by binary encoding strategy and the neighborhood search technique is employed after the iteration of GA is finished to complete the search during a very short time period. The proposed algorithm is applied to real time optimal reactive power dispatch of Hainan Power System. It is proved that with this algorithm the global optimization can be achieved and it possesses good global convergence performance. So, the presented algorithm can meet the need of real time optimal reactive power dispatching of power systems.