计算机应用与软件2017，Vol.34Issue(3)：1-6,6.DOI:10.3969/j.issn.1000-386x.2017.03.001
云环境下基于多目标优化的科学工作流数据布局策略
DATA PLACEMENT STRATEGY BASED ON MULTI-OBJECTIVE OPTIMIZATION FOR SCIENTIFIC WORKFLOWS IN CLOUD COMPUTING ENVIRONMENT
摘要
Abstract
The traditional data placement strategies for scientific workflows fail to monitor the load balancing between data centers while reducing the data transfer time.Thus, a data placement strategy based on multi-objective optimization is proposed.Firstly, the strategy generates the placement scheme of fixed datasets.Then it uses multi-objective optimization-based algorithm KnEA(Knee Point Driven Evolutionary Algorithm) to place flexible datasets, and then obtain the placement scheme of all datasets.The algorithm KnEA takes advantage of characteristic of knee points which can get good convergence comparing to other non-dominated sorting individuals, and comprehensively deals with the balance between multiple objectives.That's why the data placement strategy is able to perform well in data transferring time and load balancing.The effectiveness of the proposed method is tested by comparison with two other strategies.关键词
云计算/科学工作流/数据布局/多目标优化/负载均衡Key words
Cloud computing/Scientific workflows/Data placement/Multi-objective optimization/Load balancing分类
信息技术与安全科学引用本文复制引用
程慧敏,李学俊,吴洋,朱二周..云环境下基于多目标优化的科学工作流数据布局策略[J].计算机应用与软件,2017,34(3):1-6,6.基金项目
国家自然科学基金项目(61300169). （61300169）
