Agent-based simulation for software development processes

Tobias Ahlbrecht, Jürgen Dix, Niklas Fiekas, Jens Grabowski, Verena Herbold, Daniel Honsel, Stephan Waack, Marlon Welter

Abstract

Software development is a costly process and requires serious quality control on the management level: Managing a project with more than 10 programmers over several years is a highly nontrivial task. We are building tools for helping the manager to predict the future development of the project based on certain adjustable parameters. The main idea is to view the software process as agent-based simulation in a multiagent system (MAS). This approach requires to combine three different areas: (1) mining data and patterns from projects done in the past, (2) modeling the software development process in a multiagent environment, and (3) running the simulation on a dedicated and scalable multiagent platform.
Keywords: 
agents, simulation, software/management processes, software evolution, mining software repositories, conditional random fields
Document Type: 
Articles in Conference Proceedings
Booktitle: 
Proceedings of the 14th European Conference on Multi-Agent Systems
Series: 
EUMAS 2016
Publisher: 
Springer
Month: 
12
Year: 
2016

Main menu 2

2011 © Software Engineering For Distributed Systems Group