Automated learning of operational requirements

Automated learning of operational requirements

Abstract

Requirements Engineering involves the elicitation of high-level stakeholders goals and their refinement into operational system requirements. A key difficulty is that stakeholders typically convey their goals indirectly through intuitive narrative-style scenarios of desirable and undesirable system behaviour, whereas goal refinement methods usually require goals to be expressed declaratively using, for instance, a temporal logic. Currently, the extraction of formal requirements from scenario-based descriptions is a tedious and error-prone process that would benefit from automated tool support.

We present an ILP methodology for inferring requirements from a set of scenarios and an initial but incomplete requirements specification. The approach is based on translating the specification and scenarios into an event-based logic programming formalismand using a non-monotonic ILP system to learn a set of missing event preconditions. We then show how this learning process can be integrated with model checking to provide a general framework for the elaboration of operational requirements specifications that are complete with respect tohigh-level system goals. The contribution of this work is twofold: a novel application of ILP to requirements engineering, which demonstrates also the need for non-monotonic learning, and a novel integration of model checking and inductive logic programming.