Wei Yan

Evolving robust GP solutions for hedge fund stock selection in emerging markets

Stock selection for hedge fund portfolios is a challenging problem for Genetic Programming (GP) because the markets (the environment in which the GP solution must survive) are dynamic, unpredictable and unforgiving. How can GP be improved so that solutions are produced that are robust to non-trivial changes in the environment? We explore an approach that uses subsets of extreme environments during training.

Behavioural GP diversity for dynamic environments: an application in hedge fund investment

We present a new mechanism for preserving phenotypic behavioural diversity in a Genetic Programming application for hedge fund portfolio optimization, and provide experimental results on real-world data that indicate the importance of phenotypic behavioural diversity both in achieving higher fitness and in improving the adaptability of the GP population for continuous learning.

Special issue on Bio-inspired Learning and Intelligent Systems for Security (BLISS-07)

This special issue is based on the 2007 ECSIS Symposium on Bio-inspired Learning and Intelligent Systems for Security (BLISS-07) that was held in Edinburgh, Scotland, UK. That successful symposium emphasized reliable, versatile, and intelligent systems employed by a broad range of security applications. The goal was to integrate developers of intelligent systems with those who use them in security applications, including project managers, system integrators, and end users. As here used intelligent systems denote those artificial computational systems that operate in part or fully autonomously and that display behavior that if it were to be observed in animals, would normally become associated with intelligence of one sort or another. Systems with different degrees of autonomy of operation benefit greatly from incorporating aspects and mechanisms that are found in a broad range of biological systems, from survivability and adaptation of the simple living structures to learning, creativity, cognition and various forms of intelligence that are normally associated with humans. These features are often incorporated into algorithms by mimicking the biological processes that provide the inspiration. Such intelligent systems have been applied to a wealth of practical problems, including those in security. Examples of such applications discussed at the symposium include the detection and prevention of cybercrimes and identity theft, internet security, security of financial systems, security of public transportation systems, emergency response systems, combining space-based systems with geographical information systems, etc. A subset of selected papers presented at BLISS-07 have been chosen for this special issue of Journal of Soft Computing. These papers deal with a broad range of applications relevant to security including Voice over Internet Protocol (VOIP), identifying faces, image analysis, determining the source of an unknown airborne contaminant release, encryption for communication, among others. Two additional papers were contributed from a similar special issue being edited by Dr. Daniel Howard. All papers contained herein have relevance for security in either the civilian or defense arenas. We thank the authors who have all taken considerable effort to expound on their work for permanent archival in this journal. We similarly thank the reviewers who have facilitated this process, enhancing the quality of this special issue. We especially thank the editors of Journal of Soft Computing, particularly Brunella Gerla, for patience as we worked to make this special issue a reality. We hope the reader will capture a piece of the excitement and collegiality present in Edinburgh for the BLISS-07 conference as preserved in these papers. S. E. Haupt (&) Applied Research Laboratory, The Pennysylvania State University, State College, PA, USA e-mail: seh19@psu.edu

Diverse committees vote for dependable profits

Stock selection for hedge fund portfolios is a challenging problem for Genetic Programming (GP) because the markets (the environment in which the GP solution must survive) are dynamic, unpredictable and unforgiving. How can GP be improved so that solutions are produced that are robust to non-trivial changes in the environment? We explore an approach that uses a voting committee of GP individualswith differing phenotypic behaviour.