Application of Multi-Agent Systems on the Software Project Scheduling Plan and Control

• Main Authors: Cai, Sheng-Wei, 蔡聖偉
• Other Authors: Huang, Ming-Shang
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/81059790663938895736

碩士 === 國立屏東科技大學 === 資訊管理系所 === 97 === Effective planning and controlling is considered as an important research topic in a distributed software project organization. Therefore, developing a project scheduling and control system to support a project organization’s work such as activities, task assignments, scheduling and control, etc. should have significant impacts on project performance. Due to the changes in requirements and the need for the application of key resources for a software project, it is necessary to spend more times on conforming software requirements and the necessary key resources. To resolve the issues mentioned above, Theory of Constraints (TOC) proposes a method of applying critical chain and buffer management to deal with scheduling and control of software projects. So far, TOC is considered as an effective way to save times spending on the project. Based on the above discussions, a research framework is formulated and a model, which integrates TOC with software project scheduling and control process, is proposed. According to related research on TOC, applying TOC into the software project management has the following benefits: (1) to eliminate the occasion of Self-Fulfilling Prophecy and the Deadline Effect, (2) to release project resource to support other project work, and (3) to save a proportion of project working hours simultaneously. This research applied GAIA methodology, which was proposed by Wooldrideg, Jenning and Kinny. GAIA methodology is used to analyze decentralized project organization and design multi-agent systems in this research. It also uses JADE as the developing tool to build up multi-agent systems with software project planning and control system to support the activities of project planning and control to accomplish the project on time. In order to verify the applications in practice, we use a representative software project to demonstrate the applicability of the proposed model. A simulation study is conducted by using urgent, general, and non-urgent project cases. According the results of the study, the proposed method based on TOC has the following benefits: (1) to have more flexibility than the TOC method which uses 50-50 rules with different proportion of time-reduction, (2) in terms of software project, the proposed method is more suitable to the characteristics of software project, and to meet the user’s demand and the uncertain needs of software project requirements, (3) to keep a certain standard of quality without reducing reasonable project time, (4) to control the progress of software projects by using the buffer management, and (5) to release resources to support other project by using the critical chain of projects and buffer management to reduce project times. The findings of this study can provide useful suggestions for practitioners.