基于改进遗传算法的配电网动态无功优化

综合混沌变异的泛化能力和邻域搜索的局部寻优能力,实现配电网动态无功优化问题的快速求解.利用混沌神经元的输入输出特性建立变异算子与种群多样性测度的自适应关系,实现种群多样性的动态调节,提出基于优秀个体特征信息的邻域搜索,实现局部寻优.在编码过程中,结合配电网动态无功优化的特点提出一种由投运组数和投运时间构成的两段式整数编码方式,缩短了染色体长度、消除了不可行码.IEEE69配电网算例结果表明本文方法在全局收敛性以及克服早熟等方面具有优势.

Dynamic Reactive Power Optimization Method for Distribution Network Based on Improved Genetic Algorithm

By combining the generalization ability of chaotic mutation and the local search ability of neighborhood search method, a new dynamic reactive power optimization method for distribution network was proposed. The self-adapt relationship between mutation operator and population diversity was built based on the input-output relationship of chaotic neuron, which implemented the dynamical change of population diversity. In order to accelerate the convergence, neighborhood searching was applied on the basis of the feature information of the excellent individuals during each evolution. A two-section integer coding method which contained the capacitor number and switch time was presented. This coding method could shorten the length of chromosome and eliminate the infeasible code effectively. The proposed method was applied to IEEE 69 bus system. Optimization results have shown that it has strong global searching capability and has advantages in overcoming premature phenomena.