A systematic business model-driven approach for deriving UML-based requirement specifications

Other Title(s)

نهج منظم يعتمد على نماذج الأعمال لاشتقاق متطلبات النظام ممثلة بلغة التصميم الموحدة

Dissertant

Turkman, Salam

Thesis advisor

Tawil, Adil

Comitee Members

Jabari, Muhannad
Ilayyan, Dirar

University

Birzeit University

Faculty

Faculty of Engineering and Technology

University Country

Palestine (West Bank)

Degree

Master

Degree Date

2017

English Abstract

Requirement engineering, or elicitation as also commonly known, is a critical stage in software engineering that enables requirement engineers extract the right knowledge of the system, functional and constraints, needs of stakeholders.

This is often both an error prone and a time expensive process.

Research shows that the majority of the errors in the software functionality, often discovered in later stages of development, are directly linked to the mistakes (e.g.

misunderstanding, ambiguity/interpretations, correctness, etc.) made during the requirement elicitation phases.

Therefore, many researchers proposed different approaches to enhance the existing requirements engineering techniques to both reduce such mistakes and speed up the requirements engineering process.

One of these approaches is based on utilizing business process modelling to take benefit from business process models that are already available in an enterprise to derive requirements.

Many organizations have their existing business process models in the form of working instructions, consisting of embedded business rules and processes that provide enough valid details for specifying, important aspects of, software systems.

Upon reviewing the research done in this field, many approaches have been proposed to automatically generate requirements specifications from business models (see chapter 3), but they fail to achieve transformation without significant manual intervention.

In this thesis, we propose a structured approach to derive system requirements based on business process models (named BMSpec) that employ a set of systematic steps to improve business process models before transforming them into requirement models.

The transformation is then done by mapping objects in a business process model (represented in XML/BPMN) to a UML-based use case model and structured-text use cases descriptions.

The transformation is performed by algorithms that base their methods on a set of heuristic rules developed part of this research.

Although this transformation is not fully automated, but it aims to achieve a greater degree of automation with more precise and valid requirement specifications, thus enabling to overcome many of the problems that arise in the requirement elicitation phases, such as misunderstanding of the business by requirement engineers, lack of focus on the objective of the system, and miscommunication between stakeholders and system requirement engineer, and reducing effort, time and eventually errors in the software specifications.

37 business process models, classified into 6 levels of richness, have been used to validate and evaluate BMSpec, in terms of its efficiency, sufficiency and correctness.

The evaluation results show that BMSpec’s generated requirements were correct for the tested case studies, 90% sufficiency, of covering BPMN’s notations, and higher generation efficiency directly proportional to the level of richness of the input business process model.

Main Subjects

Information Technology and Computer Science

No. of Pages

181

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-779851