改进的模糊聚类算法在负荷预测中的应用

Application of Improving Fuzzy Clustering Algorithm to Power Load Forecasting

针对传统的c均值模糊聚类算法易陷入局部最优解、初始值c值的给定存在着很大的人为因素以及在整个计算过程中无法自我调节的缺陷,利用遗传算法的全局寻优能力并采用一种新式的双码染色体编码方法对传统的c均值模糊聚类算法进行了改进,同时将这一自适应的SFGO(SamplingFuzzyc-meanswithGeneticOptimization)算法运用到电力系统的中长期负荷预测中,得到了比较好的效果。

Classical fuzzy c means have some drawbacks such as unsatisfactory performance in finding global optimum,difficulty in choosing initial c value and lack of self-adaptive ability.To overcome these drawbacks,this paper proposes an improved fuzzy c means algorithm named self-adaptive SFGO,which takes the advantage of GA in finding global optimal solution.A new coding method called double chromosome code is used in this algorithm.Case studies verify that this algorithm can yield good forecasting results for long-term power loadforecasting problems.