• 更多栏目

    叶鑫

    • 教授     博士生导师   硕士生导师
    • 主要任职:经济管理学院院长、党委副书记
    • 其他任职:电子政务模拟仿真国家地方联合工程研究中心 副主任
    • 性别:男
    • 毕业院校:大连理工大学
    • 学位:博士
    • 所在单位:经济管理学院
    • 办公地点:经济管理学院D239
    • 电子邮箱:yexin@dlut.edu.cn

    访问量:

    开通时间:..

    最后更新时间:..

    Role network model-based support for RESTful business processes

    点击次数:

    论文类型:期刊论文

    发表时间:2014-11-01

    发表刊物:DATA & KNOWLEDGE ENGINEERING

    收录刊物:SCIE、EI

    卷号:94

    期号:PA

    页面范围:38-64

    ISSN号:0169-023X

    关键字:Workflow management; Business process; Human-intensive systems; REST; Rule-based; RNM

    摘要:The WWW is increasingly used for process-intensive business scenarios that connect services and applications together. Imperative programming and centralised traditional workflow engines lack the flexibility to support such distributed systems. RESTful business process (RESTfulBP) is an architectural style that introduces a set of constraints and uses process fragments to enable process coordination and choices at the endpoints to improve system adaptability, especially for human-intensive Web-based business processes. However, the original solution had two limitations: 1) the endpoint process fragments cannot distinguish between obligation and permission semantics to communicate richer choices of next-steps, and 2) it was time-consuming to create process fragment variants in an imperative style (either using complex workflow models or manually). In this paper, we propose a rule-based approach using the Role Network Model (RNM) to address these limitations. We introduce deontic logic into a process fragment and enable declarative generation of process fragments from rules. The proposed solution is evaluated in a real-world property valuation application using the B2B standard process from the Australian lending industry. The evaluation indicates that the new approach has improved both the design and runtime adaptability of the resulting system. Through surveys of other users employing our approach, it further demonstrates improved expressiveness, usability and general usefulness. (C) 2014 Elsevier B.V. All rights reserved.