基于模糊聚类-量子粒子群算法的用电特性识别

为解决应用传统模糊C均值（fuzzy C-means, FCM）算法进行电力负荷模式提取时存在的对初始聚类中心敏感、聚类数目不易确定等问题，构建表征聚类效果的目标函数，并针对传统智能寻优算法易收敛、陷入局部最优等缺陷，采用一种量子编码的粒子群算法进行全局寻优以确定最佳聚类中心及分类数目，在确定最佳聚类中心及聚类数目基础上，构建能够全面反映各类型负荷的特征向量，最后通过与传统FCM算法下的计算结果进行对比，验证了该方法在用电识别方面的有效性及正确性。

Electricity Characteristic Recognition Study Based on Fuzzy Clustering-Quantum Particle Swarm Algorithm

In allusion to such defects as sensitive to initial clustering center and not convenient to determine clustering number during utilizing traditional fuzzy C-Means （FCM） algorithm to extract power load patterns, this paper constructed objective function to reflect clustering effect, and used a quantum particle swarm algorithm for global optimization to determine the optimal clustering center and classification aiming at the defects of traditional intelligent optimization algorithm, such as easy convergence, falling into local optimum, etc. After determining the optimal clustering center and clustering number, the characteristics vector was constructed to fully reflect each kind of load. At last, by compared with the calculated results of traditional FCM algorithm, the effectiveness and correctness of the proposed algorithm in electricity recognition were verified.