Browsing by Author "Skolicki, Zbigniew"
Now showing 1 - 3 of 3
Results Per Page
Sort Options
Item An Analysis of Island Models in Evolutionary Computation(2007-12-19T18:54:55Z) Skolicki, Zbigniew; Skolicki, ZbigniewIsland models (IMs) are a class of distributed evolutionary algorithms (EAs) in which the population is split into multiple sub-populations called islands. Separate EAs run independently on each island, but they interact by means of migrating individuals. Therefore, IMs are different both from a single-population standard EA, as well as from a set of multiple isolated EAs. IMs are interesting for several reasons. They have been reported to yield better results than standard EAs. IMs are also advantageous when computational tasks must be distributed across multiple machines because their structure is easy to parallelize. However, despite many studies, no comprehensive theory describing their behavior has been developed. Due to the lack of theory and a complex architecture with many control parameters, setting up IMs has been a trial-and-error process, guided mostly by “rules of thumb.” In this dissertation, I adopt a two-level (intra- and inter-island) view of IMs and show how this approach makes understanding their dynamics easier. They behave very differently than standard EAs, and in order to take full advantage of this, I propose a better utilization of the inter-island level of evolution. In particular, I argue for setups with many relatively small islands, and I also show that compositional evolution may scale to the inter-island level. The two levels of evolution influence each other, and I analyze this interaction more deeply. Migrations profoundly change the local dynamics and stimulate evolution, which often ultimately results in better performance. I study the role of genetic operators in this behavior and also create mathematical models of after-migration dynamics. This analysis gives us a better understanding of mixing and the survival of genes locally, and these processes in turn determine the type and level of interaction between islands globally. Further, using island heterogeneity enhances the inter-island evolution. Following the study, I analyze IM behavior on a range of test problems, including two complex domains. This dissertation improves our understanding of the dynamics of IMs and suggests a qualitative change in the way we think about them. This perspective offers new guidelines for configuring IM parameters and opens new directions for future work.Item Intelligent agent for designing steel skeleton structures of tall buildings(2002-11) Skolicki, Zbigniew; Kicinger, Rafal P.The paper discusses a study on the application of intelligent agents (IAs) to conceptual designing. It provides an overview of the state-of-the-art in the areas of ontologies and IAs. Next, the system Disciple, a learning intelligent agent shell, and the system Inventor 2001, evolutionary design support tool, both developed at George Mason University, are briefly presented. Further, the paper introduces the developed ontology for a class of steel skeleton structures of tall buildings. This ontology was used to build an IA for the selection of initial parent design concepts in evolutionary designing. A description of the developed agent is provided as well. Finally, examples of design concepts proposed by the agent are presented. The paper also contains conclusions and recommendations for further research.Item Proactive infrastructure security: Evolutionary generation of terrorist scenarios(George Mason University Press, 2003) Arciszewski, Tomasz; De Jong, Kenneth A.; Sage, Andrew; Goode, Mike; Kicinger, Rafal P.; Skolicki, ZbigniewThe objectives of this working paper are to propose a general concept of proactive security in the context of co-evolutionary computation and to briefly discuss the initial results of research recently began. First, the paper provides an overview of infrastructure security in the context of asymmetric threats. Next, concepts of proactive security are proposed based on co-evolution of terrorist scenarios and security plans. The paper also presents an outline of generation of terrorist scenarios in the context of conceptual design. Finally, it describes TerrorMax/Capitol Hill, a demonstration system being developed for dealing with the generation of terrorist scenarios related to the Capitol Hill in Washington DC. The paper also provides initial discussions of this recently initiated project.