基于遗传算法的物理数据模型优化研究  

Research on Optimization of Physical Data Model Based on Genetic Algorithm

作　　者：杜舒明 胡璇 李情 DU Shuming;HU Xuan;LI Qing(Guangzhou Power Supply Bureau,Guangdong Grid Co.,Ltd.,Guangzhou 510610,China)

机构地区：[1]广东电网有限责任公司广州供电局,广东广州510610

出　　处：《微型电脑应用》2023年第3期108-111,共4页Microcomputer Applications

摘　　要：为了实现海量数据的高质量查询和调度,降低查询消耗、提高查询效率,研究基于遗传算法的物理数据模型优化方法。构建动态—增量物理数据模型,利用状态基态库存储当前时刻高频变化的数据对象,通过动态增量库体现存储数据变化过程,结合状态基态表和增量表建立基于优先级的历史库,提高数据的查询能力、降低对硬盘的需求;运用自适应遗传算法,经种群初始化、适应度函数设计、交叉概率和变异概率选取等步骤,对物理数据模型进行查询优化,获取最佳查询结果。实验结果表明:该方法的迭代收敛效果好,采用差异大的交叉算子和变异算子可提升物理数据模型的查询能力;查询任务越多,该方法查询消耗比越低,优势越显著;且用户并发数量增加对于该方法并发延迟时间的影响甚微。In order to achieve high quality query and scheduling of massive data,reduce query consumption and improve query efficiency,an optimization method of physical data model based on genetic algorithm is studied.A dynamic and incremental physical data model is constructed,the ground state database is used to store the data objects that change frequently at the current moment,the dynamic incremental database is used to reflect the change process of the stored data,and the historical database based on priority is established by combining the ground state table and the incremental table,so as to improve the data query ability and reduce the demand on the hard disk.Through the steps of population initialization,fitness function design,crossover probability and mutation probability selection,an adaptive genetic algorithm is used to query and optimize the physical data model to obtain the best query results.The experimental results show that the iterative convergence effect of this method is good,and the query ability of the physical data model can be improved by using the crossover operator and mutation operator with great difference.The more query tasks are,the lower the query consumption ratio is,and the more significant the advantage is.And the increase of the number of concurrent users has little effect on the concurrent delay time of this method.

关 键 词：遗传算法 物理数据模型 动态增量 交叉算子 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]