论文部分内容阅读
为了快速有效地求解电力系统无功优化问题 ,提出了一种分布式并行计算的遗传算法。它采用主从方式来组织局域网内的多台机器进行并行计算——由 1台主机进行选择和遗传操作 ,并根据负荷均衡的原则调度多台从机计算潮流以给出个体适应值。根据无功优化的特点 ,为了增加算法并行度 ,就编码方案、基于多目标函数的适应度求解和遗传操作等方面对遗传算法进行了详尽的设计。文中还着重分析了并行处理效率的相关问题。算例表明该方法不仅取得了较好的优化效果 ,而且显著地提高了计算速度
In order to solve the reactive power optimization problem in power system quickly and effectively, a distributed genetic algorithm with parallel computation is proposed. It adopts the master-slave mode to organize multiple computers in the LAN for parallel computing. One host selects and heeds the mains operation, and schedules multiple slaves to calculate the tidal current according to the principle of load balancing to give individual fitness values. According to the characteristics of reactive power optimization, in order to increase the degree of parallelism of the algorithm, the genetic algorithm is designed in detail in terms of coding scheme, fitness solution based on multi-objective function and genetic operation. The paper also focuses on the analysis of the parallel processing efficiency related issues. The example shows that this method not only achieves better optimization results but also significantly increases the computational speed