基于模糊聚类的复杂系统资源调度

Resource Scheduling of Complex System Based on Fuzzy Clustering

为了提高复杂软硬件系统的资源利用率、及时响应实时任务,采用模糊聚类方法建立了任务-资源模糊聚类分析模型.用平均距离表示模糊集的隶属度,并定义了模糊相似系数,设计了资源聚类算法(resource clustering algorithm,RCA),将历史任务-资源库里的任务与实时任务进行聚类.系统根据聚类结果分析实时任务资源占用情况,从而实现系统对实时任务的快速响应.在某复杂软硬件系统上进行了仿真测试,结果表明所提算法可实现资源的合理调度.

To improve the resource utilization of complex hardware/software systems, a fuzzy taskresource clustering model was proposed in response to real-time tasks, using the average distance to show the membership of fuzzy sets. The fuzzy similar coefficient was defined. In addition, the Resource Clustering Algorithm （RCA） was designed to cluster real-time tasks and historical tasks, and according to clustering results, the systems would analyze resources occupation and respond to real-time tasks rapidly. Through the simulation test on a certain complex hardware/software system, the result shows that the proposed algorithm can be used effectively in the resources scheduling.