基于数据生成—消耗依赖的语义工作流并行化重构方法  

Parallelism refactoring algorithm of semantic workflows based on dataproduce—consume dependency

作　　者：孙晋永[1] 闻立杰[2] 匡增雄 李涛[1] 张展 SUN Jinyong;WEN Lijie;KUANG Zengxiong;LI Tao;ZHANG Zhan(Guangxi Key Laboratory of Trusted Software,Guilin University of Electronic Technology,Guilin 541004,China;School of Software,Tsinghua University,Beijing 100084,China)

机构地区：[1]桂林电子科技大学广西可信软件重点实验室,广西桂林541004 [2]清华大学软件学院,北京100084

出　　处：《计算机集成制造系统》2020年第6期1636-1650,共15页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(61862016,61961007);广西自然科学基金资助项目(2017GXNSFAA198283,2018GXNSFAA294093,2018GXNSFAA138090);广西可信软件重点实验室资助项目(KX201723)。

摘　　要：业务过程模型的并行度是其质量的重要指标,并行度高的业务过程模型质量更好,运行效率更高。针对提高语义工作流质量的需求,以及现有业务过程模型并行化方法不能处理语义工作流中数据生成—消耗依赖、资源约束等问题,提出一种基于数据生成—消耗依赖的语义工作流并行化重构方法。首先,使用节点编号法和最近公共前驱法获取语义工作流的任务执行关系矩阵;然后,分析了任务节点间的数据生成—消耗依赖,获得数据依赖矩阵;接着,提出任务执行关系更新规则,结合数据依赖矩阵更新任务执行关系矩阵,得到基于数据依赖的任务执行关系矩阵;进一步,设计了兼顾资源约束的语义工作流重构算法以生成并行化后语义工作流;最后,提出一种语义工作流的并行度计算方法以评估并行化后语义工作流的并行程度,并开发了一个交互式的语义工作流并行化重构软件。仿真实验结果表明,所提算法提高了语义工作流的并行度,改善了语义工作流的质量,为提高基于语义工作流的业务过程运行效率提供了有效支持。The parallelism of business process models is an important indicator of its quality.The business process model with high parallelism has a better quality and a higher operating efficiency.Aiming at the requirement of improving semantic workflows qualities and the fact that existing methods of business processes model parallelism cannot deal with the data produce-consume dependency and resource constraints in semantic workflows,a parallelism refactoring algorithm of semantic workflows based on data produce-consume dependency was proposed.The task relations matrix of semantic workflow was obtained by using numbered nodes and the nearest common ancestor.The data dependency matrix was obtained by analyzing the data produce-consume dependency between task nodes.The update rules of task relations were proposed,with which the task relations matrix was updated according to the data dependency matrix to gain the task relations matrix based on data dependency.A refactoring algorithm based on the new task relations matrix was designed to generate the parallelized semantic workflow while giving consideration to resource constraints.A measure was proposed to assess the degree of parallelism for semantic workflows,while interactive software for parallelism refactoring of semantic workflows was implemented.The simulation results showed that the proposed algorithm could improve the parallelism and qualities of semantic workflows,and provide an effective support for improving the operating efficiency of business processes based on semantic workflows.

关 键 词：语义工作流 并行化 数据生成—消耗依赖 资源约束 并行度 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]