Software Development Project Management Based On System Dynamics

"Software Crisis" has plagued software industry for a long time. Lackingfundamental understanding of dynamic software development process is the majorcause of "Software Crisis". Without such an understanding the likelihood of anysignificant gain in management of software development is questionable. Theobjective of this paper is to improve understanding of general process by whichsoftware development project is managed through building an integrative systemdynamics model of software development project management. The structure of this paper is as follows: The background and purpose of the research is outlined in introduction part. In Chapter 1, the theories and models of software development projects andapplications of system dynamics in software development project are reviewed to startthe research of this paper. Chapter 2 to Chapter 4 is the core of this paper. In Chapter 2, model structure andparameterization are described. In Chapter 3, a case study is conducted to test thevalidity of the model. In Chapter 4, the model is used as an experimentation vehicle tostudy and predict the dynamics implications of an array of managerial policies andprocedures. In Chapter 5, the work of this paper is summarized,the potential of systemdynamics simulation model as a learning tool to improve software developmentproject management process and the future work is discussed. The innovations of this paper are: 1. Build an integrative system dynamics model of software development projectmanagement with explicit consideration of "soft factors" which might play asimportant roles in software development process but are often neglected in classicaltheories. The model divides software development project management activities intofour areas: software production, human resource management, planning and controland models software development project management dynamic process from astrategic view. III复旦大学博士论文 摘要 2. Unlike existing models where development stages to be modeled usuallydetermined by model structure, array application in the model of this paper makes itflexible and easy to model software development projects with different stages. 3. The model of this paper can be applied as a post-mortem evaluation tool forproject performance analysis. Model application effectively avoids subjective guess asall the results could trace their causes in the model. 4. The simulation model of this paper makes perfectly controlled experimentspossible and can be applied to quantitatively predict the implications of managerialpolicies safely and with low costs.