A Context-Aware Software Adaptation Framework

对于在运行时动态调整其行为以响应变化的软件系统的需求越来越大。对这些系统的研究是由来自两个社区的研究人员进行的，他们的研究重点不同：自适应和上下文感知。一方面，上下文感知的研究更多地关注如何建模、处理和管理上下文信息，但仅限于系统如何适应上下文信息的意外变化。另一方面，对自适应性的研究更多的是关于如何适应系统的结构和/或行为以响应需求和/或上下文变化（可能是未预料到的变化），而较少关注上下文如何建模、处理和管理。在实践中，两者之间的界限相当模糊，并且上下文感知自适应系统的工程需要以系统的方式考虑这两个方面。

There is an increasing demand for software systems that dynamically adapt their behavior at runtime in response to changes. Research into these systems has been conducted by researchers from two communities with two different emphases: self-adaptivity and context-awareness. On one hand, research in context-awareness is more concerned with how to model, process, and manage the context information but limited on how a system adapts itself in response to unanticipated changes in the context information. On the other hand, research in self-adaptivity is more about how to adapt the system's structure and/or behavior in response to requirements and/or context changes (possibly unanticipated changes), with less attention to how context is modeled, processed, and managed. In practice, the line between the two is rather blurred, and the engineering of a context-aware adaptive system needs to consider both aspects in a systematic manner. The objective of this research project is to investigate, design, and develop a novel framework to identify the requirements of considering context-awareness and self-adaptivity in an integrated manner. First, the framework provides a novel context modeling technique that captures all the system environment aspects and supports later adaptation and evolution during the software execution; second it maintains multiple system models (i.e. the structure, behavior and quality models) and their consistency at runtime, and enables system runtime evolution; and third it provides adaptation mechanisms that can cope with unanticipated environment changes using the system-context relationships. We believe that the benefit of this project to the targeted sectors of the economy is twofold. First, it provides a superior context-aware framework that will help to develop and implement sophisticated algorithms and methods for system runtime adaptivity. Second, we believe the particular modeling techniques we aim at developing for adaptive applications will help industrial companies to effectively employ methods so that the adaptive systems can be kept up-to-date during their evolution and maintenance.