Author

Bharvi Chhaya

Date of Award

12-2020

Document Type

Dissertation - Open Access

Degree Name

Doctor of Philosophy in Electrical Engineering & Computer Science

Department

College of Engineering

Committee Chair

Shafagh Jafer, Ph.D.

First Committee Member

Nickolas Macchiarella, Ph.D.

Second Committee Member

Massood Towhidnejad, Ph.D.

Third Committee Member

Keith Garfield, Ph.D.

Fourth Committee Member

Omar Ochoa, Ph.D.

Abstract

The use of domain-specific languages (DSLs) has increased manifold for problem solving in specific domain areas as they allow for a wider variety of expressions within their domain. Modeling using DSLs has shown high increases in productivity after accounting for the time and cost expended in developing them, making them a suitable target for improvement in order to reap higher rewards. The currently used approach for domain modeling involves the creation of an ontology which is then used to describe the domain model. This ontology encapsulates all domain knowledge and can be cumbersome to create, requiring external sources of information and assistance from a domain expert.

This dissertation first discusses the use and importance of DSLs for scenario generation for a domain and presents an extension to the Aviation Scenario Definition Language (ASDL). The main contribution of this dissertation is a novel framework for scenario based development of DSLs, called the Domain-Specific Scenario (DoSS) framework. This framework proposes the use of scenarios in natural language, which are currently used in requirements engineering and testing, as the basis for developing the domain model iteratively. An example of the use of this approach is provided by developing a domain model for ASDL and comparing the published model with one obtained using DoSS. This approach is supplemented with a case study to validate the claim that DoSS is easier to use by non-experts in the domain by having a user create a model and comparing it to one obtained by the author. These models were found to be almost identical, showing a promising return for this approach. The time taken and effort required to create this model by the user were recorded and found to be quite low, although no similar results have been published so no comparison could be made. State charts are then used for verification of scenarios to ensure the conformity between scenarios and models. The dissertation also discusses applications of the ideas presented here, specifically, the use of ASDL for Air Traffic Control training scenarios and the use of DoSS for ontology generation.

Share

COinS