Hybrid-Fixing:上下文一致性错误的正确修复

Hybrid-Fixing:Toward Sound Fixing of Context Inconsistency

在普适计算中,上下文持续快速变化,上下文感知应用根据上下文变化自动调整自身的行为以作出适应.然而,由于不可预测和控制的环境噪声以及环境动态变化等诸多因素的影响,环境上下文会发生一致性错误,从而导致应用表现异常甚至失效.为了解决这些问题,上下文一致性错误需要被自动并正确地修复,现基于已有工作提出了一项新的修复技术hybrid-fixing,它结合了对一致性约束的静态分析和修复动作的动态产生,即使一致性约束内部存在复杂依赖关系,也能确保所生成的修复用例必然正确.实验结果表明,这项修复技术大幅提高了一致性约束内部存在复杂依赖关系下一致性错误修复的成功率,并只花费了很小的时间开销.

In pervasive computing, environmental contexts are subject to frequent and rapid changes, and context-aware applications adapt their behavior accordingly. However, context inconsistency occurs due to various reasons including unpredictable and uncontrollable environmental noises and dynamics, which results in application anomaly or even failure. To address this problem, context inconsistency should be timely detected and then fixed in an automated and sound way. Based on our previous work we propose two techniques, named complete-fixing and CoSound-fixing, to fix context inconsistency automatically for context-aware applications. However, the two techniques are subject to some limitations in that complete-fixing has a time complexity issue and does work so efficiently, and CoSound-fixing does not have a satisfactory fixing success rate. In this paper, we propose a new fixing technique, named hybrid-fixing, which combines the static analysis of consistency constraints and the dynamic generation of repair actions to ensure the soundness of its generated repair cases, even when there exist complex dependencies inside consistency constraints. Experimental results show that our hybrid-fixing significantly increases the fixing success rate for detected context inconsistencies, as compared with CoSound-fixing when facing complex dependencies inside consistency constraints, while still incurring minor time cost only and achieving the fixing of context inconsistency in a fully automated way.