无功功率优化的改进退火选择遗传算法

作提出的改进退火选择遗传算法用于电力系统无功优化，其特点在于采用灵敏度选择无功补偿地点，缩小了遗传算法的搜索范围；用退火选择进行个体更新，保持了群的多样性，避免陷入局部最优；采用十进制编码、竞争繁殖和逐代记录最优法来加快计算速度和精度；用不定交叉和变异概率来更好地发挥交叉和变异的作用，并且进行适应度定标，促进全局最优。通过对TEEE6节点系统计算分析表明，该方法和简单遗传算法相比有更快的计算速度和更强的全局收敛能力。

A MODIFIED GENETIC ALGORITHM WITH ANNEALING SELECTION FOR REACTIVE POWER OPTIMIZATION

A modified genetic algorithm with annealing selection for reactive power optimizing is proposed. The algorithm possesses the following features: The location for reactive power compensation is selected by sensitivity, thus, the seeking range of genetic algorithm is reduced. To keep the multiform of the group the substitution of individuals is implemented by simulated annealing selection, so it is avoided to fall into local optimum. To improve the accuracy and to speed up calculation the decimal coding, tournament selection and optimal individual record are adopted. To exert the action of crossover and mutation the undefined crossover and mutation probability are used. Additionally, to promote the global optimum the adaptability scaling is performed. The result of calculation, taking the IEEE 6 bus system as an example, shows that comparing with conventional genetic algorithm, this method has faster calculation speed and better global convergence ability.