基于多目标权衡的数据网格作业调度算法

动态变化的数据网格环境下，很大一部分服务器和带宽资源于分发业务和下载类服务的传递，其资源请求任务受调度长度、时间跨度、负载均衡、服务质量、经济原则及安全性能等诸多因素制约，难以用单一的特征来表示。鉴于此，基于多目标权衡提出一种针对多业务作业请求矩阵，权衡多参数均衡的作业调度算法。仿真实验表明，该算法能在系统重载时提升系统整体的吞吐率，适应大规模图状数据处理场景.

Data Grid job scheduling algorithm based on the multi-objective trade-off

In the dynamically changing data grid environment, for a large part of the server and bandwidth resources distribution and download service delivery, the resource request tasks are restricted by scheduling length, time span, load balancing, service quality, economy and safety performance and many other factors, it is difficult to represent with an single feature. In view of this, based on multi-objective trade-off, this paper presents a job scheduling algorithm to tackle multi-service job request matrix and multi-parametric equalization. Simulation results show that this algorithm can significantly improve overall system throughput and adapt to the large-scale map-like data processing scene.