Context-aware Service Modeling And Conflicts Discovery Based On Petri Net

Services in a pervasive computing environment is transparent and able to be accessed anywhere at any time. This makes pervasive computing a main trend in the development of computer science and its applications. As the key technology of pervasive computing, context-aware computing provides users with task relevant information and services by sensing and reasoning the contexts and judging user behaviors accordingly. In a context-aware system, services are plentiful and diversified to meet different user needs. Thus the relationships among services become complicated and the possibility of service conflicts increases. Service conflicts can be detected and coordinated during the system run time by using dynamic monitoring mechanisms. However, this may overload the system and degrade system performance. And there is also a possibility that services may be delayed because of the coordination at run time. Therefore, a context-aware service modeling and simulating methodology is needed to simulate the practical situation as much as possible and to find out hidden conflicts before the system go live.As a graphical and mathematical modeling tool that is applicable to multiple kinds of systems, Petri net has been widely applied in context-aware service modeling researches because of its capability in describing dynamic system behaviors. However, the basic Petri net is not fully capable of modeling the characteristics of context-aware services. Therefore lots of researchers have extended it to be more suitable for the modeling of context-aware service and its environment. But still a well-defined Petri net is lacked to be able to neatly model the dynamic, complex and random contexts as well as the complicated service constraints. In this thesis, we defined a Context-aware Service Model based on Petri net (CSMP) which is able to describe the above features and proposed a context-aware service modeling method as well as a service conflict discovery method based on it. First, we give the detailed definition of CSMP. In CSMP, the diversity of contexts is modeled by token with colors, and the complexity and randomness of contexts is modeled by timestamps and random functions. The transition priority and ConDec constraints also make it possible to describe multiple service constraints. Then we studied the properties of CSMP model. As most of the context-aware service logics are described by ECA rules, we also present four basic transformation patterns that transform ECA rules into CSMP models. In the core of this thesis, we elaborated the CSMP based context-aware services and their environment modeling method systematically and studied the conflicts discovery method from the system constraint perspective. Finally, we verified the effectiveness of this method by studying an intelligent assistant system for Alzheimer’s patients with CPN Tools.The result of this thesis shows that CSMP is able to model context-aware system precisely and hidden service conflicts are able to be found based on this method. This method can be used to support system design and service logic design. It also can be used to support system developers to find a way to improve system and reduce problems during system run time. The case study result indicates that this method has applicable values and worth for the further research.