基于时间事件因果关系检测的工业软件的需求确认方法

Checking of Timed Casual Relation of Events for Requirements Validation of Industrial Software

工业软件深度参与研发设计、生产制造、运营管理和维护服务等方面,软件的行为符合业务的需要至关重要。因此,工业软件的开发需要进行需求确认,即确认系统的行为满足利益相关者(应用方的操作人员,通常是生产和运维中涉及的各方面的工作人员)的要求。业务方面,利益相关者的期望通常表现为关心的事件间的因果关系。针对工业软件的时间融合于行为、复杂度高、规模大等特点,提出一种基于时间事件因果关系检测的需求确认方法,检测用UML/MARTE+CCSL模型表达的系统行为是否满足相应的时间事件因果关系。包括:定义时间事件因果关系表达利益相关者的期望;抽取模型的多图协作下的系统整体行为生成CCSL(clock constraint specification language)规约;结合模型检测技术和社区发现算法检测该行为规约是否满足时间事件因果关系。通过比较实验评估了方法的有效性和实用性,特别是引入社区发现算法处理规模大、复杂度高的规约效果显著。

Industrial software is essential in product design,development,maintenance,and services.It is crucial to ensure that the software behaviors meet specific business needs.Therefore,industrial software development necessitates requirements validation,explicitly confirming that system behavior meets stakeholdersexpectations(operators of applied companies).These expectations are often in the form of event casual relations with timing constraints.Industrial software is time-critical,large and highly complex.Considering these features,we proposed an approach to check event-causal relations with timing constraints on UML/MARTE+CCSL models for requirements validation of industrial software.Event casual relations with timing constraints are defined for expressing stakeholders’expectations;CCSL(clock constraint specification language)specification is built for capturing the overall behaviors under the cooperation of multi-diagrams in the model;community detection algorithm is integrated with model checking techniques to accomplish the checking of CCSL specification against timed event-causal relations.The effectiveness and practicability of our approach are illustrated by comparison experiments,especially the benefits of dealing with large scale and high complexity by bringing in a community detection algorithm.