基于粒子群聚类算法的大坝安全监控模型

Dam safety monitoring model based on PSO-fuzzy clustering algorithm

将粒子群算法与模糊聚类算法相结合,建立了基于粒子群聚类算法的大坝安全监控模型.该算法将分类矩阵作为粒子的编码形式,依据粒子的个体极值和全局极值,充分利用正反馈计算信息,自适应性地确定模糊分类矩阵和聚类中心.工程算例表明:粒子群聚类算法进一步提高了聚类算法的区间预报能力;对于高维优化问题,粒子的搜索过程比较复杂,该算法的收敛速度较慢.

A dam safety monitoring model was presented in this paper by combining a fuzzy clustering algorithm and a particle swarm optimization （PSO） algorithm. Based on the position vector of particles represented by a classification matrix, the individual extremum and global extremum of each particle and the positive feedback information in the PSO, the fuzzy classification matrix and clustering center were adaptively determined. The result of an engineering application shows that, compared with the traditional fuzzy clustering algorithm, the PSO-fuzzy clustering algorithm improves the clustering effect and interval forecasting ability. It is also concluded that for high-dimension optimization problems, the convergence speed of this algorithm is slow because of its complex search process.