Automated Error-Detection and Repair for Compositional Software Specifications

摘要: The complexity of error diagnosis in requirements specifications, already high, is increased when refer to various system components, on whose interaction the system’s aims depend. Further, finding causes error, and ways overcoming them, cannot easily be achieved without a systematic methodology. This has led researchers explore combined use verification machine-learning support automated software analysis repair. However, existing approaches have been limited by using formalisms which modularity compositionality explicitly expressed. In this paper we overcome limitation. We define translation from representative process algebra, Finite State Processes, into action language \(\mathcal{C}+\). enables forms not supported previous methods. then logic-programming equivalent \(\mathcal{C}+\), apply inductive logic programming for learning repairs components while ensuring no new errors are introduced interactions with other maintained. These two phases iterated until correct specification reached, enabling rigorous scalable repair component-based specifications.