【作者】 刘昕； 李清； 陈禹六；

【Author】 LIU Xin, LI Qing, CHEN Yuliu (Department of Automation, Tsinghua University, Beijing 100084, China)

【机构】 清华大学自动化系；

【摘要】 过程集成不仅要对经营过程进行描述、优化和设计,还包括实施、运行和监控。但是,现实中经营过程设计能力与实际性能不一致的现象使得我们无法正确估计和预测经营过程性能参数, 妨碍了过程集成的有效实现。本文结合优化设计、方案选优、运行调度等方面的需求,分析了客户需求多变的环境下经营过程管理(BPM)所面临的如何描述和估计TO-BE经营过程性能的问题,引入了经营过程动态性能的概念,并借鉴学习曲线的研究成果,提出了基于学习曲线的经营过程动态性能监测和预测神经网算法,并进行了实例验证。更多还原

【Abstract】 Process integration covers not only the description, optimization and design of Business Process(BP), but also the implementation, operation and monitoring of BP. But, the difference between design performance and practical performance of BP prevents us from monitoring and prediction, which impedes the implementation of process integration. In this paper, the problem that Business Process Management(BPM) encounters are examined with the requirements of BP optimization and operation schedule, and the concept of BP dynamic performance is introduced. Then a learning-curve-based business process dynamic performance monitoring and prediction neutral network algorithm is presented, and a case study is performed.更多还原