基于遗传算法的分布式数据库数据分配策略研究

摘要

In order to improve further the accessing efficiency of DDBS,this paper analyzes the data allocation problem and builds a corresponding mathematical model in consideration of practical application.Based on the existing research of solving this problem with genetic algorithm,new measures are adopted to improve the genetic algorithm,for examples,an excellent original population is created by selective breeding in the initialization phase,the individuals are chosen by a rule which combines order selection and elitist selection during the evolutionary process,the big mutation is utilized to broaden the searching extent of the algorithm,and so on.Then a new data allocation strategy is put forward.The simulation results show that this strategy can be used to get a better data allocation scheme compared with other methods.%为进一步提升分布式数据库系统的访问效率,对数据分配问题进行分析,并结合实际应用构建出相应的数学模型.在现有以遗传算法求解该问题的研究基础上,重新优化了遗传算法,如采取选育措施得到优良初始种群、以顺序选择和最佳保存相结合的手段进行个体选择、利用大变异操作拓宽算法搜索广度等,进而提出了一种新的数据分配策略.仿真结果表明,与其他分配方法相比,运用该策略能够求得更为理想的数据分配方案.