Scalability Optimization Algorithm of Power Big Data Storage Architecture Based on Improved Bayesian Network
Affected by the traditional local search and optimization mechanism of Bayesian networks,in the expansion and opti-mization of power big data storage frameworks,it is impossible to balance the dependency relationship among global extended data,resulting in local deviations in optimization and severe performance degradation of the extended storage part.In order to solve this problem,a Bayesian network for power big data storage architecture is constructed.The expandable nodes of the power big data storage architecture are established and the upper limit of the original framework nodes is determined.The scal-ability coding calculation of the improved Bayesian network structure and the fitness evaluation of the storage framework scal-ability optimization are redefined as the optimization algorithm.The data from simulation and comparative experiments indicate that the proposed algorithm has both read response and write response indicators controlled below 60 ms,and performs best in controlling fluctuation amplitude.For sample sets with different sizes,the mean deviations of read and write performance are 0.2 and 0.3 respectively,which are close to the ideal state.The maximum fitness coefficient value is 2.The proposed algo-rithm is practical and feasible,with higher practical application value and market promotion value.
improved Bayesian networkpower big datastorage frameworkoptimization algorithm
万方数据
维普
