Motivation
Current Software Engineering (SE) practice is a form of human-led search for solutions which meet needs and constraints under limited resources. Often there will
be uncertainty or conflict, both between and within functional and non-functional criteria. A belief shared by a great part of the community is that, as systems
get bigger, more distributed, more dynamic and more critical, this labour-intensive search will eventually hit fundamental limits. In other words, it will not be
feasible to continue to develop, operate and maintain software systems in the traditional way, without (partly) automatically supporting the search for solutions.
Computational Intelligence (CI) techniques have a track record of success in other engineering disciplines, characterised by large search spaces, complex and
conflicting constraints, or uncertainty on information. With this motivation, in the last few years, there has been an increasing interest in the use of CI for
solving SE problems.
On the one hand, recently, a major research effort is being devoted to the use of search methods in SE, obtaining encouraging outcomes. However, although the scope
is wide, most of the works to date have focused on classical Evolutionary Algorithms, and on software testing as the application area.
On the other hand, Artificial Neural Networks and Fuzzy Logic, together with Machine Learning based approaches, have gradually been applied to SE, covering issues
like software development effort estimation, defect prediction models, software modules classification, design and reuse, among many others.
This special session has two main goals: first, to review the latest state-of-art on CI for SE and, second, to bring together researchers to exchange ideas and discuss
the common challenges faced when solving SE problems with CI techniques.
|
Scope
We invite authors to submit original and unpublished works showing the use of Metaheuristics and CI methods for solving SE problems. Not only empirical works, but also
theoretical contributions are welcome. Regarding Metaheuristics, we encourage methods not previously applied to SE. Machine Learning related approaches are considered
as well.
Methodological fields of interest include, but are not limited to:
- Metaheuristics:
- Evolutionary Algorithms
- Bio-Inspired methods: Particle Swarm Optimisation, Estimation of Distribution Algorithms, Ant Colony, etc.
- non-Evolutionary Computation procedures: Simulated Annealing, Tabu Search, Path Relinking, Scatter Search, Cross-Entropy, etc.
- Other computational paradigms:
- Artificial Neural Networks: kernel approaches, Support Vector Machines, mixture models, ensembles, etc.
- Fuzzy Logic: Fuzzy Sets approaches, Fuzzy Rule Based Systems, Evolutionary Fuzzy Systems, etc.
- Hybrid Intelligent Systems
Challenging application issues include, but are not restricted to:
- Software project management: planning, software release, effort estimation, etc.
- Software quality assurance: quality modelling, reliability estimation and prediction, etc.
- Software testing, verification and validation: fault identification, test data generation, bug fixing, etc.
- Software design: modularisation, re-factoring, metrics, object oriented software, reuse, etc.
- Non-functional SE: scalability, robustness, software understanding, self-improving/self-changing elements, etc.
|
