Using ontologies is a promising way to automate some time-consuming activitiesin software development processes, such as requirements analysis and validation,and software verification. Ontologies can also be used to support software developmentprocesses by sharing domain knowledge between development phases. Systemspecification ontologies (SSOs) can be developed from software requirementsspecifications (SRS) to represent requirements and knowledge about the softwaresystem that is being developed. SSOs can be interpreted by a tool to infer the knowledgethat can be used to automate some time-consuming activities in software developmentprocesses, in order to reduce the cost and time of the software development.However, manually developing SSOs to represent the software requirementsknowledge is a time-consuming and a challenging task.
The focus of this PhD project is to understand how to generate system specificationontologies semi-automatically from software requirements specifications. This thesiswill guide how software requirements specifications can be processed to capturedomain knowledge and knowledge about the software system, and represent softwarerequirements in an ontology semi-automatically. We will develop a frameworkand a set of methods within this framework which can be a possible solution tocreate semi-automatically ontologies from SRS. The constructed ontologies can beused in software processes to automate time-consuming activities, such as requirementsvalidation and software test case generations. The developed framework andmethods will be evaluated by comparing with existing ontology learning methodsand by using constructed ontologies to automate test case generation which is partof software testing process. The proposed solution will be evaluated by in differentcase studies that can have similar software requirements specifications.
Research proposal, PhD programme, University of Skövde