الفهرس 
Introduction
Organization of the ThesisOnly 14 pages are availabe for public view
 |  | from | 114 |  |  |
| | | from | 114 | | |
Abstract
Recently, it has been recorded that the majority of industrial software companies are increasingly dealing with complicated systems of high number of requirements, time limit, and high expectations in terms of efficiency and quality of the resulting software and conflicting objective. However, most software system developments are observed to be kind of iterative and incremental processes that are hardly ever completed in one single release. On the other hand, it is challenging when it comes to select the requirements from a large number of candidates to achieve the overall system goal as well as goals of multiple customers, each of whom may have competing and frequently conflicting priorities.
This work then comes to present, in view of optimization problem, the problem of software engineering in general, and requirements engineering in particular. On one hand, Requirements engineering represents the first step of software development process as it is one of the main concerns of software engineers. System requirements selection, on the other hand, is the engineering process to select an optimal set of system requirements for implementation in the consecutive stage of software system building process. This process in carried based on many requirements proposed by the customers on condition that budget and customer satisfactions are being effectively balanced. Consequently, this NP-hard problem is an important issue, involving several conflicting objectives that have to be processed by software companies when developing new software systems. Software systems have to perform their function within resource constraints, but they also have to cover the largest number of customer requirements. Additionally, in real cases of the problem, the requirements selection process highly suffers from interactions and other constraints which leading to intensively complicate this problem.
In this thesis, in the sake of finding optimal solution, meta-heuristic technique is presented along with adapted/modified multi-objective genetic algorithm which has been successfully applied on several real cases of the above mentioned problem. The system requirements selection problem is formulated as a multi-objective optimization problem with two objectives. These objectives are: minimizing the total system’s development cost and maximizing customer’s satisfaction totality. Moreover, the method is non-dominated sorting genetic algorithm with pareto tournament for multi-objective optimization (NSGA-IIPT) has been used to solve real cases of the problem and tested with case studies on two real datasets to demonstrate and prove the efficiency of proposed approach. Experimental results show that the NSGA-IIPT algorithm can efficiently
generate high-quality solutions and performs better than the other pertinent previously published algorithms in the literature under a set of public datasets. These methods were evaluated against each other by comparing their results in terms of multi-objective metrics.