Topic：Theme-based Release Planning

Speaker: Dr. Guenther Ruhe(University of Calgary)

Time: 10:30am, 6th June 2014

Venue: Lecture Room (334), 3rd Floor, Building #5（计算机科学重点实验室报告厅）

Speaker Bio：

Dr. Ruhe received a doctorate rer. nat degree in Mathematics with emphasis on Operations Research fromFreiberg University, Germany and a doctorate habil. nat. degree from both the Technical University ofLeipzig and University of Kaiserslautern. He had a visiting professorship at University of Bayreuth in1991/92 and got an Alexander von Humboldt research fellowship at University of Kaiserslautern in 1992.

Ruhe was visiting scientist at the IBM Research Centre in Heidelberg in 1993. From 1996 until 2001 he wasdeputy director of the Fraunhofer Institute for Experimental Software Engineering Fh IESE, Germanywww.iese.fhg.de.

He holds an Industrial Research Chair in Software Engineering at University of Calgary. This is a jointposition between department of Computer Science and department of Electrical and Computer Engineering.Dr. Ruhe is proud to lead a research team whose primary aim is to establish and demonstrate scientificexcellence in the area of software engineering decision support and to collaborate with industry. The area ofresearch is highly relevant, because software and systems decisions are of paramount importance for successin business today. He was a Principal Investigator for various European research projects (CEMP, PROFES, HYPER, SOFTQAULI, KOBRA, CORONET). He has extensive experience in industry collaboration andtechnology transfer (Allianz, Bosch, City of Calgary, Corel, DaimlerChrysler, Ericsson, Nokia, Nortel,Schlumberger, Siemens, Stonesoft, Trema).

Abstract：

Release planning is a mandatory part of incremental and iterative software development. For the decision about which features should be implemented next, the values of features need to be balanced with the effort and readiness of their implementation. As goal, traditional planning looks at the sum of the values of individual and potentially isolated features. As an alternative idea, a theme is a meta-functionality which integrates a number of individual features under a joint umbrella. That way, possible value synergies from offering features in conjunction (theme-related) can be utilized.

Themes might be defined, manually, upfront or as the result of computer-based analysis. In this paper, we propose an analytical approach to detect themes from a given set of feature dependencies. On top of an existing release planning methodology called EVOLVE II, our approach applies clustering performed on a feature dependency graph. The release plans generated from such an approach are a balance between two goals: (i) considering the values of individual features, (ii) detecting and utilizing synergy effects between semantically related features.

We also model theme-based release planning as a bi-objective (search-based) optimization problem. Trade-off solutions are balancing their preference between individual and theme-based planning objectives. We apply a two-staged solution approach. In Phase 1, the existing Non-dominated Sorting Genetic Algorithm-II (NSGA-II) is adapted. Subsequently, the problem of guiding the user in the set of non-dominated solutions is addressed in Phase 2. We propose and explore two alternative ways to select among the (potentially large number) of Pareto solutions. The applicability and empirical analysis of the proposed approach is evaluated for two explorative case study projects having 50 resp. 25 features grouped around 8 resp. 5 themes.