基于约束满足的大数据聚类中心调度算法仿真

针对传统调度方法聚类中心不受限,导致数据迁移次数以及能耗高的问题,提出基于约束满足的大数据聚类中心调度算法。运用转换方法,列出大数据中存在一定关系的存储块相关序列。参照数据聚类偏差点,确定存储中序列中心点,汇总序列偏向判断大数据存储序列的中心区域,在区域中选取密度近似为0的中心点,通过预测中心方进行阶梯式探测,得出存储中心,作为约束条件,采用2-opt算法将聚类中心之间距离加入到新生成序列中,调换初始序列和发生变化的序列新变量值,列出大数据预测模型逐步递进方程,使用最小二乘计算方法计算最小参数值,动向预测部分、单向的大数据,满足大数据聚类中心实施调度,实现调度最大化。通过仿真表明,所提方法在合理运用大数据的同时,能有效地控制调度风险、降低迁移次数、减少调度能源消耗。

Simulation of Large Data Clustering Center Scheduling Algorithms Based on Constraint Satisfaction

In this article,a big data clustering center scheduling algorithm based on constraint satisfaction was presented.Firstly,the conversion method was used to list the memory block sequences in big data.According to the clustering deviation point,the center point of sequence could be determined.Based on the sequence bias,it was able to judge the central region of big data storage sequence.In this region,the center point whose density was close to zero was chosen,and then stepwise detection was carried out through the prediction center,so that the storage center was found.On this basis,the 2-opt algorithm was used to add the distance between cluster centers to new sequence.The initial sequence and new variable value of changed sequence were swapped,and then the progressive equations in big data prediction model were listed.The least square method was used to calculate the minimum parameter,the trend prediction and the one-way big data,so as to meet the scheduling implementation of big data cluster center.Thus,the scheduling maximization was achieved.Simulation results show that the proposed method can effectively control the scheduling risk,decrease the number of migration and reduce the scheduling energy consumption when u?sing the big data reasonably.