Date of Award

7-2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Engineering and Sciences

First Advisor

Marco Carvalho

Second Advisor

Muzaffar Shaikh

Third Advisor

Philip Bernhard

Fourth Advisor

Walter Bond

Abstract

Development of multi-agent systems is negatively impacted by the lack of process standardization across the major development phases, such as the requirements analysis phase. This issue creates a key barrier for agent technology stakeholders regarding comprehending and analyzing complexity associated with agent-oriented specification. Instead, such fundamental low-level infrastructure is loosely attended to in an ad-hoc fashion, and important aspects of requirements analysis are often neglected altogether. The IEEE Std 830 model is a recommended practice aimed at describing how to write better quality requirements specification of conventional software. Knowing that agent-based computing is a natural evolution of the conventional approaches to software development, we believe that the requirements phase in multi-agent systems can benefit from applying the IEEE Std 830 model which results in high-quality and more accepted artifacts. This PhD research provides a criteria-based evaluation derived from the software engineering body of knowledge guide to assessing the adoption degree of agentoriented methodologies to software requirement standards. It also proposes a model-driven approach to restructuring and extending the IEEE Std 830 model to become more suitable to specify requirements of multi-agent systems. The updated model is subjected to several different types of evaluations to ensure its applicability and usefulness.

Download

Share

COinS