Guest editorial searchbased software engineering

Authors: Gordon Fraser Jerffeson Teixeira de Souza

Publish Date: 2014/09/07

Volume: , Issue: , Pages: 1-3

PDF Link

Abstract

Searchbased software engineering SBSE models complex software engineering problems as optimization problems and uses metaheuristic search algorithms to solve these SBSE has the potential to address the challenges posed by the growing size of complexity of modern computer systems and has already been successfully applied to solve problems in nearly all software development life cycle phases This special issue presents three articles demonstrating the versatility of SBSE by tackling three very different aspects of software engineering Software module clustering network optimisation requirements for dynamic adaptive systemsWith modern computer systems growing in size and complexity the demands to functionality scalability and robustness of software engineering techniques present ever new challenges Searchbased software engineering SBSE has emerged as a promising direction of research that can successfully counter some of these challengesSBSE involves the automatic – or semiautomatic resolution of complex software engineering problems modeled as optimization problems using search algorithms in particular metaheuristics An exponential growth of the number of published papers in this area in the last decade bears witness to the popularity searchbased approaches within the software engineering research community and SBSE has been successfully applied to solve problems in nearly all software development life cycle phasesThe paper “An Experimental Evaluation of the Importance of Randomness in Hill Climbing Searches applied to Software Engineering Problems” by Marcio Barros addresses a fundamental cross cutting concern related to the importance of randomness in SBSE A range of experiments analyzes whether quasirandom sequences could be employed as an alternative to classic pseudorandom number generators Contrary to past experiments in this domain the author finds no evidence that quasirandom sequences would outperform pseudorandom sequencesEric Fredericks Byron DeVries and Betty Cheng in their paper “AutoRELAX Automatically RELAXing a Goal Model to Address Uncertainty” address the problem of specifying requirements for dynamic adaptive systems such as for example robotic systems where there is uncertainty about the environment during requirements elicitation The article proposes and evaluates an approach based on searchtechniques to generate goal models based on such uncertain environmental conditionsFinally the paper “Evolving Robust Networks for SystemsofSystems is it Viable for Large Networks” by Jonathan Aitken Rob Alexander Tim Kelly and Simon Poulding looks at software systems using adhoc networks and applies search techniques to optimize these networks Initial empirical evidence is encouraging and the tool implementing the technique can produce solutions that are better than a human engineer could come up with

Keywords: