With computers becoming increasingly smaller and the number of transistors on a CPU becoming consequently increasingly higher, the speed of computers is increasing at an almost exponential rate. This and almost an equal improvement in storage capacity, coupled with the availability of more and more data resulting from decentralised locations through computer networks (especially the Internet) or from advanced measurement devices, has lead to a wealth of information hitherto not captured or unavailable.
This scenario has added a new term, “big data”, to the vocabulary of not only the scientific community working in data-driven areas such astronomy and physics, bioinformatics, computational biology, engineering (e.g. electricity networks, communication networks, image and video processing, sensor networks or the remote control of machines), etc., but also of economists, administrators, managers, etc.
Big data usually implies a collection of huge as well as diverse data sets. These no doubt provide huge opportunities in the aforementioned areas. However, opportunities always give rise to challenges too.
Big data brings with it very vital and useful information which would otherwise remain elusive and which hence, in many cases, proves to be a booster for the successful completion of scientific, engineering or business projects, increasing productivity in organisations, etc. However, this data often becomes too large and complex to capture, process, store, search, share, transfer and analyse. Traditional database management or data processing techniques fail to handle such a deluge of data for tasks such as capture and analysis. In other words, there is an imperative need to use innovative and new tools and techniques for handling big data.
In particular, from the perspective of data analytics, making sense out of big data accurately, efficiently and quickly has become a popular research challenge. Traditional data analytics require the full set of data to be bounded and available. These methods may fall short when the big data potentially amounts to infinity. The development of new techniques such as, for instance, in-memory types of analysis and incremental methods have received much attention recently from researchers.
Metaheuristics are known to be strong in optimisation when the problem is computationally difficult or simply too large. Although metaheuristics often do not result in an optimal solution, they may provide reasonable solutions within acceptable computation times, e.g. by using stochastic mechanisms. Both metaheuristics and big data analytics share a common ground of looking for approximate results out of a potentially intractable search space, via incremental operations. There has been much research effort emerging recently in applying metaheuristics to non-stationary data and incremental analytics.
This special issue aims to capture some of these efforts and so that they can contribute towards further analysis and progress. Papers are invited which look at the issues of applied metaheuristics in big data scenarios.
The objective of the issue is to bring together researchers from (i) various areas where big data problems arise and (ii) from the field of metaheuristics. Up to now there is no common forum for this research direction and publications on using metaheuristics in a big data context are widely distributed over different types of publications. This issue will provide new research results in this area and thus foster attention on the usefulness and importance of using metaheuristics to make big data problems more tractable. Applications of metaheuristics in diverse areas such as the sciences, engineering and business will surely expand the application domain of existing methods. It is also expected that significant methodological progress in metaheuristics can be stimulated because most metaheuristics applications today focus on smaller types of problems, so the suitability and effectiveness of the methods for larger problems is not obvious.
This special issue will commemorate the 5th anniversary of the International Neural Network Society (INNS) India Regional Chapter. More details of INNS-India are available here. Although this special issue is meant for commemorating this anniversary, it is open to submissions from all interested authors, and hopes to encompass every corner of the scientific world.
Suitable topics include, but are not limited to, the following:
- Decision support systems
- Swarm intelligence and optimisation
- Cuckoo search
- Bee algorithms
- Firefly algorithms
- Data mining
- Routing and scheduling optimisation
- Ant colony optimisation
- Genetic algorithms
- Big data analytics
- Intelligent information technology
- Intelligent agents and nature-inspired computing
- Large-scale logistics and production planning
- Simulation and modelling
- Text analysis and mining
- Computational biology
- Web mining
- Impacts, issues and challenges of big data - real-world applications
Deadline for manuscripts: 31 May, 2015
Notification of authors after reviews: 31 October, 2015