基于分布式K-means算法的水电厂光纤测温系统可扩展性优化

Scalability optimization of optical fiber temperature measurement system in hydropower plant based on distributed K-means algorithm

针对水电厂光纤测温系统并行扩展性和问题扩展性较差的问题,设计了基于分布式K-means算法的水电厂光纤测温系统可扩展性优化方法。根据系统真实的执行时间和能耗,计算系统通信和能耗的实际加速比,构建系统可扩展模型;利用基于分布式K-means算法的协同过滤模型分散数据集,聚合相似度高的样本数据,完成系统的可扩展性优化。实验结果表明,设计方法的性能明显优于其他方法,解决了系统并行扩展性和问题扩展性的优化问题。

Aiming at the parallel scalability of optical fiber temperature measuring system in hydropower plant,the scalability optimization method based on distributed K-means algorithm is designed.According to the real execution time and energy consumption of the system,calculate the actual acceleration ratio of system communication and energy consumption,constructed the system scalable model;The data set was dispersed and the sample data with high similarity was aggregated by the collaborative filtering model based on the distributed K-means algorithm,to complete the scalability optimization of the system.The experimental results show thatthe design method performs significantly better than other methods,solving the optimization problems of the system parallel scalability and the problem scalability.