Please use this identifier to cite or link to this item: http://prr.hec.gov.pk/jspui/handle/123456789/15512
Title: Intelligent Recommender and Decision Support System for Effective Management of Software Projects
Authors: Hamid, Muhammad
Keywords: Physical Sciences
Computer & IT
Issue Date: 2019
Publisher: National College of Business Administration & Economics, Lahore.
Abstract: Software Project Management (SPM) is a vital concern for software industries to follow best practices for successful project completion. Despite the rich availability of SPM literature, every year around 70% of projects cannot gain successful completion worldwide. Software failure influences the software industry in terms of reduced revenue, stressed teams and reduced motivation. This ultimately leads to the reduction of jobs followed by showing down the country economy in form of reduced software exports. In this regard, the three main objectives of the study were defined. The first objective of this study was to explore the literature on SPM for identifying major critical / contributing factors in software failure. The second objective of this study was to assess the impact of critical factors in the success and failure of software projects in Pakistan. The third objective of this study was the proposed Intelligent Recommender and Decision Support System (IRDSs) based on the finding of the survey. The Design Science Research (DSR) methodology was adopted to achieve the objectives of this research. The research activities were divided into three phases: in the first identification phase, it explored research articles, books, and reports dealing with software project failure through Systematic Literature Review (SLR). In the initial stage, 2079 studies were extracted using search strings. Through title base screening, 1232 studies were eliminated, reducing the remaining number to 847. Then the abstract based screening was performed on the remaining 847 papers and 791 more papers were excluded, reducing the final count to 56 papers. On these 56 papers, the snowballing method was applied and 92 more papers were identified. The inclusion and exclusion criteria were applied on these newly identified papers, where duplicated, title based screening and abstract based screening left 12 papers to be included, making a total of (56+12) 68 papers for a thorough investigation. The analysis of 68 selected research papers highlighted 13 influencing factors to software project failure, with four major, five significant and four insignificant factors. Where the major factors are incorrect cost and time estimation, scope creeping, inadequate human resources, and risk management. The analysis included 35.29% empirical studies, 47.06% general literature review and 17.65% case studies. The analysis also reflected that 86.77% of papers only examined the state of the art while only 13.23% of research studies discussed some algorithms. Further, the analysis found that only 4.41%, studies developed some automation tool and while 95.59% of the studies did not develop any tool. In the prototype design and development phase, the questionnaire design and data collection were performed. The primary purpose of this survey was to identify the main reasons behind software project failure and success in Pakistan. To accomplish this task, data were collected from 339 senior developers and project managers working in software companies throughout Pakistan with the help of the x Software Export Board (PSEB). The logistic regression enquiry was executed to measure the relationship between various factors affecting software. The results reflect that improper planning along with the wrong cost and time estimation are positively and significantly associated with software failure. The Structural Equation Modelling (SEM) was used to analyze the constructs and to assess the relationship between factors affecting software success. The empirical results showed that improper planning, inadequate human resources, wrong estimation of time and cost have a significantly negative impact on the success of software projects. Further, based on the limitation identified in the survey results, the conceptual model IRDSs was proposed. The conceptual model indicated that the scrum master is the main user of the proposed approach. The scrum master initiates a request that describes business domain requirements related to which they estimate. These requests are forwarded to the system that activate and assign the work to the relevant agents. Agents start analyzing the knowledge base as per the scrum master requirements. Once all the agents accomplish their assigned goals, their work compiled together with the estimate agent in the form of an estimate and finally, it presents to the scrum master as per their respective requests. Further, the formal specification of IRDSs is also performed using the formalism known as Z. The proposed algorithm explains various steps involved in the estimation of projects in a scrum environment. In the prototype demonstration and evaluation phase, the IRDSs system was developed and it became activated by the scrum master’s request for the allocation of human resources and estimation of time and cost for upcoming single or multiple user stories. The proposed system has two main parts. Firstly (decision support system), it gives the decision to software companies or scrum master regarding how much time and the cost is required for completing single or multiple user stories. The intelligent recommender is the second part of the system. It helps the software companies or scrum master regarding how much appropriate team members and their seniority level required to develop a particular user story or stories to complete on optimizing cost and time. An experiment on fifteen projects was performed to check the validation of the proposed approach, with Delphi and planning poker estimation methods. The overall results by applying Mean Magnitude Relative Error (MMRE), Percentage Relative Error Deviation (PRED) as evaluation measures indicated that the proposed method produced better estimation results than planning poker and Delphi methods. This research adds new knowledge from an underdeveloped country, which can open new dimensions for the IT industry and project managers working under similar circumstances.
Gov't Doc #: 20619
URI: http://prr.hec.gov.pk/jspui/handle/123456789/15512
Appears in Collections:PhD Thesis of All Public / Private Sector Universities / DAIs.

Files in This Item:
File Description SizeFormat 
Muhammad hamid 2019 CS ncbae lhr prr.pdfphd.Thesis3.63 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.