Agile Development Workflow Design Based On Petri Net

In order to respond quickly to the changes and uncertainties in business and internal process, more and more software development teams adopt agile software development methods as their model, which is characterized as incremental and evolutionary. In essence, agile development is undertaken on simple framework to manage or adapt to changes from external and internal events in controlled manners where self-organization, cooperation and co-decision, feedback loops, time-based together with emphasis on adaptability are particularly addressed. Workflow modeling and management method could analyze, control and manage the process of agile development to reduce risk and make it "lightweight". Feedback loop system of agile development workflow, agile story estimation workflow and tasks scheduling problems are proposed and discussed in this paper on the basis of theoretical research and case studies (Scrum) exploiting stochastic Petri nets(SPN), generalized stochastic Petri nets (GSPN) and genetic algorithm theories with PIPE (Platform Independent Petri Net Editor) as the simulation tool. In this paper, the contributions and academic innovations are as follows:(1) Consider agile development workflow as an adaptive control system. Proposed an open loop agile development workflow system employing GSPN based on case studies and agile development framework. Based on open-loop model, several feedback arcs are designed to integrate critical information and resource into workflow, making it a close loop workflow control system which could well behave with feedback, coordination with progress checkpoints and adaptation.(2) Developed SPN-based model and performance analysis of agile story estimation based on defining parameters and factors influencing the decision-making process. Optimizing strategies and suggestions are given with regard to self-organization management, planning meeting schedule and arrangement and better decision-making manner for agile practicers.(3) Developed time optimization and robust optimization scheduling strategies based on genetic algorithm by converting task scheduling problem in agile development workflow into resource constrained project scheduling problem. Particularly, relative amount of contribution from resources are adapted rather than absolute amount to meet the demand of free coordination under the agile development values. A simulation instance is illustrated to specify two strategies and to verify the enforceability of robust optimization scheduling strategies. An exception handling model is put forward and analyzed with Petri Nets with the aim to illustrate how uncertainties in development workflow can be controlled.In the last chapter, a summary of conclusion and prospects for further work is proposed.