遗传策略的综合改进及其在负荷建模中的应用

研究了遗传操作和控制参数选择对遗传算法性能的影响,设计了解群选择的随机-精英策略、避免近亲繁殖的双断点交叉策略和交叉与变异概率的自适应调整策略,提出了一种综合改进型遗传算法并成功地应用于基于实测参数的负荷建模。算例表明,该算法改善了进化过程中的种群多样性和早熟现象,对加速收敛、缩短辨识时间、克服模型参数分散性、提高辨识精度均有显著作用,适用于负荷建模。

Synthetic Improvements of Genetic Strategies and Their Application in Power Load Modeling

The influence of genetic operations and selection of control parameters on the performance of genetic algorithm is systematically studied. A random-elite strategy to choose colony, a two-point crossover strategy to avoid inbreeding and a self-adaptive modulating strategy to control both crossover and mutation probability are elaborately designed. A synthetically improved genetic algorithm is proposed and successfully applied to load modeling based on measured data. Practical modeling process shows that the proposed synthetically improved genetic algorithm can effectively ameliorate population diversity and overcome precocity in evolution course, and play outstanding role in accelerating convergence, shortening identification time, overcoming dispersivity of model parameters and improving identification precision. So it is suitable for load modeling.