Self-repairing Of Socio-technical Systems In Pervasive Computing Environment

For traditional software systems such as information systems, runtime adaptation only involves reconfigurations of internal elements like system parameters and runtime architectures. While socio-technical system under pervasive computing environment includes not only software but also hardware (such as sensors) and human beings as its system elements. The open and dynamic characteristic of socio-technical system will lead to uncertainty of whole system behaviors, thus often face the risk of system failure. These socio-technical systems are expected to perform in a broad set of environments with different capabilities and resources. Application requirements may change dynamically requiring flexible adaptation. Therefore socio-technical system under open, dynamic pervasive environment faces the main problem is how to implement the self-repairing capability based on complex social interactive between related software, hardware and human to reduce the incidence of system failure.To solve this problem, this paper proposes a self-repairing method based on previous work for socio-technical system under pervasive computing environment at runtime. This method is based on requirement goal model supporting multi-agent, and through applying goal model and state machine in pervasive computing system and improve context-aware modeling algorithm, event mapping rule to support such system. First, this paper proposes a context semantic modeling and inferring method. This method could use rule-based inferring to find potential system failure and provide basis for self-repairing decision based on the context monitoring at runtime. Then, to meet socio-technical system self-repairing requirement at runtime, this paper proposed a service search and retrieval algorithm based on semantic inferring and provides correspondent service quality evaluation to make the self-repairing process dynamically select and evoke proper service. Finally, to solve conflicts under multiple users, this paper defines a multiple objective conflict model, and utilize the context-aware model and conflict model to establish corresponding strategy about conflicts detection and self-repairing decision. The expectation result is that the system could resolve the conflicts and make total satisfaction highest.To demonstrate availability and effectiveness, this paper simulates the true scenario in our daily life. The experiment, which is based on Java and Arduino kit, provides a design structure about socio-technical system under pervasive computing environment, and describes the main module and their functions. Preliminary experimental results of case studies verify the validity of the method:This method could precisely analyze user behavior and reason about the sensors needed and retrieve proper services to help users meet their goals based on uncertain behaviors. In the meantime, when multiple users meet conflictions, the system could give a solution that makes the total satisfaction high under the certain time.