Research on Power Quality Data Placement Strategy Based on Clustering-Genetic Algorithm

Data placement in multiple data centers is the key to optimizing computing performance. The power quality monitoring system of State Grid uses data centers of all provinces to store monitoring data nearby, and realize data calculation across provinces by workflow. This type of placement based on the area has the problem that the data transmission time is too long. In view of the above problem, a data placement strategy based on clustering-genetic algorithm is proposed. In this algorithm, the initial population generation algorithm based on data correlation clustering makes the individuals in the initial population have a certain degree of clustering accuracy and strong diversity. The initial population generation algorithm and the adaptive operator of genetic algorithm cooperate with each other, which effectively enhances the algorithm's ability to find the best. Compared with the traditional data placement strategy, the experimental results show that the data placement strategy based on clustering-genetic algorithm is more efficient than the traditional genetic algorithm strategy. Compared with the traditional clustering algorithm, it can effectively reduce the data transmission time of the workflow runtime, the average reduction of about 15%.