@inproceedings{antonisse.keller:dynamic1986, Address = {Ft. Worth, TX}, Author = {Jim Antonisse and K. S. Keller}, Booktitle = {Proceedings of the 7th Annual Digital Avionics Systems Conference}, Title = {Dynamic Evaluation of Imprecisely Specified Knowledge}, Year = 1986 } @inproceedings{antonisse.keller:dynamic1988, Address = {John Hopkins University Applied Physics Laboratory, Laurel, MD}, Author = {Jim Antonisse and K. S. Keller}, Booktitle = {Proceedings of the 1988 Data Fusion Symposium}, Title = {Dynamic Evaluation of Sources in Rule-Based Sensor Fusion Systems}, Year = 1988 } @inproceedings{antonisse.keller:genetic1987, Address = {Cambridge, MA}, Author = {Jim Antonisse and K. S. Keller}, Booktitle = {Proceedings of the Second International Conference on Genetic Algorithms}, Title = {Genetic Operations for High-Level Knowledge Representations}, Year = 1987 } @inproceedings{antonisse:a-grammar-based1990, Address = {Indiana University}, Author = {Jim Antonisse}, Booktitle = {Foundations of the Genetic Algorithm Workshop}, Title = {A Grammar-Based Genetic Algorithm}, Year = 1990 } @inproceedings{antonisse:a-new-interpretation1989, Address = {George Mason University, Fairfax, VA}, Author = {Jim Antonisse}, Booktitle = {Proceedings of the Third International Conference on Genetic Algorithms}, Title = {A New Interpretation of Schema Notation that Overturns the Binary Encoding Constraint}, Year = 1989 } @inproceedings{antonisse:unsupervised1990, Author = {Jim Antonisse}, Booktitle = {IEEE Systems, Man and Cybernetics}, Month = {November}, Title = {Unsupervised Sensor Credit Assignment in Knowledge-Based Sensor-Fusion Systems}, Year = 1990 } @inproceedings{arciszewski.de-jong.ea:proactive2003, Abstract = {The 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.}, Address = {Fairfax, VA}, Author = {Arciszewski, Tomasz and De Jong, Kenneth A. and Sage, Andrew and Goode, Mike and Kicinger, Rafal and Skolicki, Zbigniew}, Booktitle = {Proceedings of the Workshop on the Critical Infrastructure Protection Project, Airlie Center, Warrenton, VA, August, 2003}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/ArciszewskiCIPP2003.pdf}, Editor = {Woodcock, Alexander and Thomas, Kevin}, Keywords = {infrastructure security and evolutionary computation and proactive security and coevolutionary algorithms and terrorist scenarios}, Pages = {378-391}, Publisher = {George Mason University Press}, Title = {Proactive infrastructure security: evolutionary generation of terrorist scenarios}, Year = 2003 } @inproceedings{arciszewski.jong.ea:inventive1999, Author = {Tomasz Arciszewski and Kenneth De Jong and H. Vyas}, Booktitle = {Proceedings of the Fifth International Conference on the Applications of AI to Civil and Structural Engineering}, Title = {Inventive Design in Structural Engineering: Evolutionary Computing Approach}, Year = 1999 } @incollection{arciszewski.jong:evolutionary2001, Abstract = {This paper provides an overview of the state of the art of evolutionary computation in civil engineering. First, it discusses the fundamentals of evolutionary computation in the context of a unified approach recently proposed by one of the authors. Second, it discusses the main research directions, including morphogenic design and co-evolutionary design. Next, it provides a description of the emerging evolutionary computation analysis, including an analysis in the context of complex adaptive systems and one based on the visualization of the gene pool and of the fitness function. Finally, the paper contains the conclusions and recommendations for the further research.}, Author = {Tomasz Arciszewski and Kenneth De Jong}, Booktitle = {Civil and Structural Engineering Computing}, Pages = {161 -- 184}, Publisher = {Saxe-Coburg Publications}, Title = {Evolutionary Computation in Civil Engineering: Research Frontiers}, Year = 2001 } @inproceedings{arciszewski.kicinger:proactive2005, Abstract = {The main objective of the paper is to propose several novel approaches to security of complex infrastructure systems, which can be utilized in the development of a class of computer tools for infrastructure protection. First, the paper introduces the concept of proactive infrastructure security and compares it with reactive security. The comparison is done in the context of the generation and evaluation of both the terrorist and security scenarios, which are also introduced and described. Next, the paper discusses both the evolutionary and co-evolutionary generation of terrorist and security scenarios and discusses various computer tools, which have been developed at George Mason University for infrastructure protection. Finally, the paper briefly overviews the concept of cellular automata and proposes how cellular automata could be used in the development of computer tools for infrastructure protection. The paper ends with the initial research conclusions and various suggestions for further research.}, Address = {Boston, Massachusetts, USA}, Author = {Arciszewski, Tomasz and Kicinger, Rafal}, Booktitle = {Working Together: Conference on Public/Private R&D Partnerships in Homeland Security}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/ArciszewskiDHS2005.pdf}, Keywords = {infrastructure security and critical infrastructure protection and cellular automata and evolutionary computation and coevolutionary algorithms}, Pages = {(poster)}, Publisher = {DHS}, Title = {Proactive security: From evolutionary approaches to cellular automata}, Year = 2005 } @incollection{arciszewski.kicinger:structural2005, Abstract = {This paper explores the state of the art in structural design inspired by nature and proposes an improved understanding of this emerging paradigm and its major components. First, it introduces and discusses the nature of three categories of inspiration, including visual, conceptual, and computational inspiration. Next, several important inspiration sources are identified and briefly described, including Evolutionary Computation, Coevolutionary Computation, Cellular Automata, and TRIZ. In particular, design generation mechanisms based on cellular automata are introduced with some details. They are inspired by the processes of morphogenesis occurring in nature and have great potential to generate novel designs. In this case, the design generation mechanisms are encoded in so-called generative representations, which are also described. Three major design objectives are introduced and discussed, namely optimality, creativity, and robustness. They are related to the sources of inspiration and the corresponding computational mechanisms inspired by nature. Also, three levels of integration of computational mechanisms inspired by nature are proposed and their relationship to design objectives is discussed. Finally, a general design situation, when inspiration by nature is considered, is introduced and its unified description is proposed as the first step in the direction of building a unified approach to structural design inspired by nature. The paper provides initial conclusions and discusses the most promising directions for future research.}, Address = {Stirling, Scotland}, Author = {Arciszewski, Tomasz and Kicinger, Rafal}, Booktitle = {Innovation in Civil and Structural Engineering Computing}, Download1 = {Link to the book,http://saxe-coburg.co.uk/pubs/descrip/sl2005.htm}, Editor = {Topping, Barry H. V.}, Isbn = 1874672245, Keywords = {creative design and morphogenesis and morphogenic evolutionary design and evolutionary computation and design inspired by nature and bio-inspired design and structural design and evolutionary design and coevolutionary design and generative representations and cellular automata}, Pages = {25-48}, Publisher = {Saxe-Coburg Publications}, Title = {Structural design inspired by nature}, Year = 2005 } @incollection{arciszewski.skolicki.ea:intelligent2005, Author = {Tomasz Arciszewski and Zbigniew Skolicki and Kenneth De Jong}, Booktitle = {Agents and Multi-Agent Systems in Construction}, Editor = {C. J. Anumba and O. O. Ugwu and Z. Ren}, Pages = {6 -- 30}, Publisher = {Taylor and Francis}, Title = {Intelligent Agent Fundaments}, Year = 2005 } @inproceedings{bala.de-jong.ea:hybrid1995, Address = {Montreal, Quebec, Canada}, Author = {Jerzy Bala and De Jong, Kenneth A. and J. Haung and Haleh Vafaie and H. Wechsler}, Booktitle = {Proceedings of the 14th International Joint Conference on Artificial Intelligence}, Download1 = {PDF,papers/ijcai95.pdf}, Download2 = {PostScript,papers/ijcai95.ps}, Download3 = {GZipped PostScript,papers/ijcai95.ps.gz}, Month = {August}, Title = {Hybrid Learning Using Genetic Algorithms and Decision Trees for Pattern Classification}, Year = 1995 } @inproceedings{bala.de-jong.ea:visual1996, Address = {Vienna, Austria}, Author = {Jerzy Bala and De Jong, Kenneth A. and Jeffrey Huang and Haleh Vafaie and H. Wechsler}, Booktitle = {Proceedings of the 13th International Conference on Pattern Recognition}, Publisher = {IEEE Computer Society Press}, Title = {Visual Routine for Eye Detection Using Hybrid Genetic Architectures}, Year = 1996 } @article{bala.jong.ea:using1996, Author = {Jerzy Bala and Kenneth A. De Jong and Jeffrey Huang and Haleh Vafaie and H. Wechsler}, Journal = {Evolutionary Computation}, Note = {Special Issue of Evolutionary Computation - Evolution, Learning, and Instinct: 100 Years of the Baldwin Effect}, Pages = {297--311}, Publisher = {MIT press}, Title = {Using Learning to Facilitate the Evolution of Features for Recognizing Visual Concepts}, Year = 1996 } @inproceedings{balan.luke:a-demonstration2004, Abstract = {Genetic programming may be seen as a recent incarnation of a long-held goal in evolutionary computation: to develop actual computational devices through evolutionary search. Genetic programming is particularly attractive because of the generality of its application, but it has rarely been used in environments requiring iteration, recursion, or internal state. In this paper we investigate a version of genetic programming developed originally by Astro Teller called neural programming. Neural programming has a cyclic graph representation which lends itself naturally to implicit internal state and recurrence, but previously has been used primarily for problems which do not need these features. In this paper we show a successful application of neural programming to various partially observable Markov decision processes, originally developed for the learning classifier system community, and which require the use of internal state and iteration.}, Author = {Gabriel Catalin Balan and Sean Luke}, Booktitle = {Proceedings of Genetic and Evolutionary Computation Conference}, Pages = {422 -- 433}, Title = {A Demonstration of Neural Programming Applied to Non-Markovian Problems}, Year = 2004 } @inproceedings{balan.luke:history-based2006, Abstract = {What if traffic lights gave you a break after you've spent a long time waiting in traffic elsewhere? In this paper we examine a variety of multi-agent traffic light controllers which consider vehicles' past stopped-at-red histories. For example, a controller might distribute credits to cars as they wait and award the green light to lanes with the most credits, allowing cars to keep the credits they accumulate during travel. Such history-based controllers are intended to provide a kind of global fairness, reducing the variance in mean time spent waiting at lights during trips. We compare these controllers against other multi-agent controllers which only consider present information, and discover, among other things, that while the history-based controllers are among the most robust, they often unexpectedly provide more efficiency than fairness.}, Address = {Hakodate, Japan}, Author = {Gabriel Balan and Sean Luke}, Booktitle = {Proceedings of Autonomous Agents and Multiagent Systems}, Month = {May}, Title = {History-Based Traffic Control}, Year = 2006 } @techreport{balan.richards.ea:algorithms2008, Abstract = {Solutions to non-cooperative multiagent systems often require achieving a joint policy which is as fair to all parties as possible. There are a variety of methods for determining the fairest such joint policy. One approach, min fairness, finds the policy which maximizes the minimum average reward given to any agent. We focus on an extension, leximin fairness, which breaks ties among candidate policies by choosing the one which maximizes the second-to-minimum average reward, then the third-to-minimum average reward, and so on. This method has a number of advantages over others in the literature, but has so far been little-used because of the computational cost in employing it to find the fairest policy. In this paper we propose a linear programming based algorithm for computing leximin fairness in repeated games which has a polynomial time complexity given certain reasonable assumptions.}, Author = {Gabriel Balan and Dana Richards and Sean Luke}, Institution = {George Mason University}, Number = {GMU-CS-TR-2008-1}, Title = {Algorithms for Leximin-Optimal Fair Policies in Repeated Games}, Year = 2008, Download1 = {PDF, http://cs.gmu.edu/~tr-admin/papers/GMU-CS-TR-2008-1.pdf} } @inproceedings{balan.richards.ea:long-term2008, Abstract = {How does one repeatedly choose actions so as to be fairest to the multiple beneficiaries of those actions? We examine approaches to discovering sequences of actions for which the worst-off beneficiaries are treated maximally well, then secondarily the second-worst-off, and so on. We formulate the problem for the situation where the sequence of action choices continues forever; this problem may be reduced to a set of linear programs. We then extend the problem to situations where the game ends at some unknown finite time in the future. We demonstrate that an optimal solution is NP-hard, and present two good approximation algorithms.}, Address = {Chicago, Illinois}, Author = {Gabriel Balan and Dana Richards and Sean Luke}, Booktitle = {Proceedings of AAAI Advances in Preference Workshop}, Pages = {7-12}, Title = {Long-term Fairness with Bounded Worst-case Losses}, Year = 2008 } @inproceedings{bassett.de-jong:evolving2000, Abstract = {A good deal of progress has been made in the past few years in the design and implementation of control programs for autonomous agents. A natural extension of this work is to consider solving difficult tasks with teams of cooperating agents. Our interest in this area is motivated in part by our involvement in a Navy-sponsored micro air vehicle (MAV) project in which the goal is to solve difficult surveillance tasks using a large team of small inexpensive autonomous air vehicles rather than a few expensive piloted vehicles. Our approach to developing control programs for this MAVs is to use evolutionary computation techniques to evolve behavioral rule sets. In this paper we describe our architecture for achieving this, and we present some of our initial results.}, Address = {Berlin Heidleberg}, Author = {Bassett, Jeffrey K. and De Jong, Kenneth A.}, Booktitle = {Proceedings of the Twelfth International Symposium on Methodologies for Intelligent Systems}, Download1 = {PDF,papers/Bassett00ismis.pdf}, Download2 = {PostScript,papers/Bassett00ismis.ps}, Download3 = {GZipped PostScript,papers/Bassett00ismis.ps.gz}, Editor = {Z. Ras and S. Ohsuga}, Pages = {157--165}, Publisher = {Springer}, Series = {LNAI}, Title = {Evolving Behaviors for Cooperating Agents}, Volume = 1932, Year = 2000 } @inproceedings{bassett.potter.ea:applying2005, Abstract = {Several researchers have used Price's equation (from biology theory literature) to analyze the various components of an Evolutionary Algorithm (EA) while it is running, giving insights into the components contributions and interactions. While their results are interesting, they are also limited by the fact that Price's equation was designed to work with the averages of population fitness. The EA practitioner, on the other hand, is typically interested in the best individuals in the population, not the average. In this paper we introduce an approach to using Price's equation which instead calculates the upper tails of population distributions. By applying Price's equation to EAs that use survival selection instead of parent selection, this information is calculated automatically.}, Author = {Jeffrey K. Bassett and Mitchell A. Potter and De Jong, Kenneth A.}, Booktitle = {Proceedings of Genetic and Evolutionary Computation Conference -- GECCO-2005}, Download1 = {PDF,papers/bassett05applying.pdf}, Pages = {1371--1378}, Publisher = {ACM Press}, Title = {Applying Price's Equation to Survival Selection}, Year = 2005 } @inproceedings{bassett.potter.ea:looking2004, Abstract = {In this paper we show how tools based on extensions of Price's equation allow us to look inside production-level EAs to see how selection, representation, and reproductive operators interact with each other, and how these interactions affect EA performance. With such tools it is possible to understand at a deeper level how existing EAs work as well as provide support for making better design decisions involving new EC applications.}, Author = {Jeffrey K. Bassett and Mitchell A. Potter and De Jong, Kenneth A.}, Booktitle = {Genetic and Evolutionary Computation Conference -- GECCO-2004}, Download1 = {PDF,papers/BassettPotterDejong2004gecco.pdf}, Publisher = {Springer}, Title = {Looking Under the EA Hood with Price's Equation}, Year = 2004 } @inproceedings{bassett.wiegand.ea:evolving2001, Author = {Bassett, Jeffrey K. and Wiegand, R. Paul and De Jong, Kenneth A.}, Booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) 2001}, Pages = {1133 (Poster)}, Publisher = {Morgan Kaufmann Publishers}, Title = {Evolving Multi--Agent Behaviors Using a Tunable Problem Landscape}, Year = 2001 } @mastersthesis{bassett:a-study2002, Abstract = {Generalization is an important aspect of all machine learning algorithms. Without it, learning can only occur in the simplest of problem domains. The most interesting domains tend to be more complex though. As we scale-up our algorithms to work in these complex domains, an understanding of the different generalization techniques available and the trade-offs between them becomes increasingly important. This thesis is primarily concerned with rule learning using evolutionary algorithms. We examine two commonly used generalization techniques called wildcards and partial matching, as well as a third which is a hybrid of these two. We demonstrate that the wildcards are more effective at generalizing in some domains, and partial matching is more effective in others. It is our hypothesis that the hybrid will be the more robust of the three techniques. In other words, the hybrid will generalize as well, or almost as well, as the better of the other two in a variety of domains. Two very different domains were chosen as testbeds for our experiments. The first is a concept learning domain, which tends to favor wildcards. The second is a multi-agent robotics task, where partial matching is more effective. When the hybrid was tested in these environments, the results show that it was either equivalent or superior to the other two techniques, thus verifying our hypothesis. Finally we attempt to find an explanation for these results that will help predict behavior in other domains. We examine how these generalization techniques alter the inductive bias of the learning algorithm. The analysis demonstrates weaknesses in both wildcards and partial matching. It also suggests that the weaknesses of each technique is offset by the strength of the other, thus allowing the hybrid to be more effective. An attempt was made to verify the inductive bias explanation. We use this knowledge to devise a concept learning problem which will favor partial matching instead of wildcards. But the results of this experiment show that the most accurate classifiers were produced using the wildcard generalization mechanism. The inductive bias hypothesis is clearly helpful for understanding some of the behaviors observed in the learning algorithm. None the less, further study will be required to understand a problem as complex as this one fully.}, Address = {Fairfax VA, USA}, Author = {Jeffrey K. Bassett}, Download1 = {PDF,papers/Bassett02thesis.pdf}, Download2 = {PostScript,papers/Bassett02thesis.ps}, Download3 = {GZipped PostScript,papers/Bassett02thesis.ps.gz}, School = {George Mason University}, Title = {A Study of Generalization Techniques in Evolutionary Rule Learning}, Year = 2002 } @inbook{bigbee.cioffi-revilla.ea:agent-based2007, Abstract = {The purpose of this research was to replicate the Sugarscape model (Eptstein and Axtell 1996) and simulation outcomes as described in Growing Artificial Societies (GAS). Sugarscape is a classic agent-based model and contemporary simulation toolkits usually only have a very simple replication of a few core rules. There is scant evidence of significant replication of the rules and simulation outcomes; code supplied with Repast, Swarm, and NetLogo implement a minority of the rules in Sugarscape. In particular, the standard Repast distribution only implements Growback, Movement, and Replacement. Sugarscape implementations in these toolkits are clearly provided only as basic demonstrations of how wellknown social models might be implemented, rather than complete achievements of scientific replication.}, Author = {Anthony Bigbee and Claudio Cioffi-Revilla and Sean Luke}, Chapter = {Replication of Sugarscape using MASON}, Pages = {183--190}, Publisher = {Springer}, Title = {Agent-Based Approaches in Economic and Social Complex Systems IV}, Volume = 3, Year = 2007 } @inproceedings{campbell.wu.ea:emperical2006, Abstract = {The goal of this research is to explore the effects of social interactions between individual autonomous vehicles (AVs) in various problem scenarios. We will take a look at one way to construct the social relationships and generate data from computer simulations to compare the behaviors of each. A difference can be noticed when Synthetic Social Structures (SSS) are used to control the interactions between neighboring AVs. Our experiments show that SSSs can be used to improve team performance on a problem in which a team of AVs must maneuver through a narrow corridor to reach a goal.}, Author = {Adam Campbell and Annie S. Wu and Keith Garfield and Randall Shumaker and Sean Luke and Kenneth A. De Jong}, Booktitle = {Proceedings of International Conference on Networking, Sensing, and Control}, Pages = {440 -- 445}, Title = {Emperical Study on the Effects of Synthetic Social Structures on Teams of Autonomous Vehiles}, Year = 2006 } @inproceedings{cervone.panait.ea:an-application2004, Abstract = {Multi sensor remote sensing provides real time high resolution data that can be used to study anomalous changes on land, in the ocean, and in the atmosphere associated with an impending earthquake. Anomalous behaviour in Surface Latent Heat Flux (SLHF) prior to large coastal earthquakes has been recently found. However, an SLHF time series usually contains several sharp peaks that may be associated either with earthquakes or with atmospheric perturbations. In this paper we have used evolutionary algorithms to perform a search in a large space bounded by longitude, latitude and time, to distinguish between signals associated with earthquakes and those associated with atmospheric phenomena. The algorithm finds paths which delimit the extent of the detected anomalies by optimizing an objective function that takes into consideration several aspects, such as spatial and time continuity, the magnitude of the anomalies, and the distance to the continental boundary. This search strategy is crucial for the development of a fully automated early warning system for providing information about impending earthquakes in a seismically active coastal region. Experiments have been performed over a 2000 km2 area comprising a part of the continental boundary between the African and Eurasian plate, roughly corresponding to Italy and Greece, one of the most seismically active regions. Using a 365-days-long time series, we identified three signals associated with seismic events. Additionally, it was possible to establish that the extent of the signal does not propagate further than 600 km from the epicenter of the earthquake.}, Author = {Guido Cervone and Liviu Panait and Ramesh Singh and Sean Luke}, Booktitle = {Late Breaking Papers of the Genetic and Evolutionary Computation Conference -- GECCO-2004}, Publisher = {Springer}, Title = {An Application of Evolutionary Algorithms to Study the Extent of SLHF Anomaly Associated with Coastal Earthquakes}, Year = 2004 } @inproceedings{cioffi-revilla.luke.ea:computational2006, Abstract = {We present a new international project to develop temporally and spatially calibrated agent-based models of the rise and fall of polities in Inner Asia (Central Eurasia) in the past 5,000 years. Gaps in theory, data, and computational models for explaining long-term sociopolitical change—both growth and decay—motivate this project. We expect three contributions: (1) new theoreticallygrounded simulation models validated and calibrated by the best available data; (2) a new long-term cross-cultural database with several data sets; and (3) new conceptual, theoretical, and methodological contributions for understanding social complexity and long-term change and adaptation in real and artificial societies. Our theoretical framework is based on explaining sociopolitical evolution by the process of "canonical variation".}, Author = {Claudio Cioffi-Revilla and Sean Luke and Dawn C. Parker and J. D. Rogers and W. W. Fitzhugh and W. Honeychurch and B. Frohlich and P. DePriest and Chanag Amartuvshin}, Booktitle = {Proceedings of First World Congress on Social Simulation}, Title = {Computational Modeling Frontiers in International Politics: Agent-based Modeling and Simulation of Adaptive Behavior and Long-Term Change in Inner Asia}, Year = 2006 } @inproceedings{cioffi-revilla.paus.ea:mnemonic2004, Abstract = {How does group memory affect sociality? Most computational multi-agent social simulation models are designed with agents lacking explicit internal information-processing structure in terms of basic cognitive elements. In particular, memory is usually not explicitly modeled. We present initial results from a new prototype called ``Wetlands'', designed to investigate the effect of group memory structures and interaction situations on emergent patterns of sociality or collective intentionality. Specifically, we report on initial computational experiments conducted on culturally-differentiated agents endowed with finite and degradable memory that simulate bounded mnemonic function and forgetfulness. Our main initial findings are that memory capacity and engram retention both promote sociality among groups, probably as nonlinear (inverse) functions. Wetlands 1.1 is implemented in the new MASON 3 (Multi- Agent Simulator of Networks and Neighborhoods) computational environment developed at George Mason University.}, Author = {Claudio Cioffi-Revilla and Sean Paus and Sean Luke and James Olds and Jason Thomas}, Booktitle = {Proceedings of the Conference on Colective Intentionality IV}, Download1 = {PDF,http://cs.gmu.edu/~eclab/projects/mason/publications/siena.pdf}, Title = {Mnemonic Structure and Sociality: A Computational Agent-Based Simulation Model}, Year = 2004 } @techreport{coletti.lash.ea:a-preliminary1999, Author = {Mark Coletti and Tom Lash and Craig Mandsager and R. Michalski}, Download1 = {PostScript,http://www.lychnobite.org/me/LEMmlil.ps}, Download2 = {GZipped PostScript,http://www.lychnobite.org/me/LEMmlil.ps.gz}, Institution = {George Mason University}, Number = {MLI 99-5}, Title = {A Preliminary Experimental Application of Learnable Evolution Model and Evolutionary Algorithms to Parameter Estimation in Non-linear Digital Signal Filters Design}, Year = 1999 } @inproceedings{de-jong.potter.ea:using1997, Address = {Michigan State University, East Lansing, MI}, Author = {De Jong, Kenneth A. and Mitchell A. Potter and William M. Spears}, Booktitle = {Proceedings of the Seventh International Conference on Genetic Algorithms}, Download1 = {PDF,http://www.cs.uwyo.edu/~wspears/papers/icga97.pdf}, Month = {July}, Pages = {338--345}, Publisher = {Morgan Kaufmann}, Title = {Using Problem Generators to Explore the Effects of Epistasis}, Year = 1997 } @inproceedings{de-jong.potter:complex1995, Address = {San Diego, CA}, Author = {De Jong, Kenneth A. and Mitchell A. Potter}, Booktitle = {Proceedings of the Fourth Annual Conference on Evolutionary Programming}, Download1 = {PostScript,http://www.cs.gmu.edu/~mpotter/pubs/ep95.ps}, Title = {Complex Structures via Cooperative Coevolution}, Year = 1995 } @inproceedings{de-jong.sarma:generation1992, Address = {Vail, CO}, Author = {De Jong, Kenneth A. and Jayshree Sarma}, Booktitle = {Foundations of Genetic Algorithms - 2}, Download1 = {PDF,papers/gengaps.pdf}, Download2 = {PostScript,papers/gengaps.ps}, Download3 = {GZipped PostScript,papers/gengaps.ps.gz}, Editor = {Darrell Whitley}, Pages = {19--28}, Publisher = {Morgan Kaufmann}, Title = {Generation Gaps Revisited}, Year = 1992 } @inproceedings{de-jong.sarma:on-decentralizing1995, Address = {Pittsburgh, PA}, Author = {De Jong, Kenneth A. and Jayshree Sarma}, Booktitle = {Genetic Algorithms: Proceedings of the Sixth International Conference (ICGA95)}, Download1 = {PDF,papers/sarma-icga95.pdf}, Download2 = {PostScript,papers/sarma-icga95.ps}, Download3 = {GZipped PostScript,papers/sarma-icga95.ps.gz}, Month = {July}, Pages = {17--23}, Publisher = {Morgan Kaufmann}, Title = {On Decentralizing Selection Algorithms}, Year = 1995 } @inproceedings{de-jong.schultz:using1988, Address = {Ann Arbor, MI}, Author = {De Jong, Kenneth A. and Alan C. Schultz}, Booktitle = {Proceedings of the Fifth International Machine Learning Conference}, Download1 = {GZipped PostScript,ftp://ftp.aic.nrl.navy.mil/pub/schultz/papers/gina.ps.gz}, Pages = {284--290}, Title = {Using Experience-Based Learning in Game Playing}, Year = 1988 } @article{de-jong.spears.ea:using1993, Abstract = {In this paper, we explore the use of genetic algorithms (GAs) as a key element in the design and implementation of robust concept learning systems. We describe and evaluate a GA-based system called GABIL that continually learns and refines concept classification rules from its interaction with the environment. The use of GAs is motivated by recent studies showing the effects of various forms of bias built into different concept learning systems, resulting in systems that perform well on certain concept classes (generally, those well matched to the biases) and poorly on others. By incorporating a GA as the underlying adaptive search mechanism, we are able to construct a concept learning system that has a simple, unified architecture with several important features. First, the system is surprisingly robust even with minimal bias. Second, the system can be easily extended to incorporate traditional forms of bias found in other concept learning systems. Finally, the architecture of the system encourages explicit representation of such biases and, as a result, provides for an important additional feature: the ability to dynamically adjust system bias. The viability of this approach is illustrated by comparing the performance of GABIL with that of four other more traditional concept learners (AQ14, C4.5, ID5R, and IACL) on a variety of target concepts. We conclude with some observations about the merits of this approach and about possible extensions.}, Author = {De Jong, Kenneth A. and William M. Spears and Diana F. Gordon}, Download1 = {GZipped PostScript,http://www.aic.nrl.navy.mil/~spears/papers/mlj93.ps.gz}, Editor = {J. Grefenstette}, Journal = {Machine Learning}, Note = {Special issue on genetic algorithms}, Pages = {161--188}, Publisher = {Kluwer Academic}, Title = {Using Genetic Algorithms for Concept Learning}, Volume = 13, Year = 1993 } @inproceedings{de-jong.spears.ea:using1994, Abstract = {Our theoretical understanding of the properties of genetic algorithms (GAs) being used for function optimization (GAFOs) is not as strong as we would like. Traditional schema analysis provides some first order insights, but doesn't capture the non-linear dynamics of the GA search process very well. Markov chain theory has been used primarily for steady state analysis of GAs. In this paper we explore the use of transient Markov chain analysis to model and understand the behavior of finite population GAFOs observed while in transition to steady states. This approach appears to provide new insights into the circumstances under which GAFOs will (will not) perform well. Some preliminary results are presented and an initial evaluation of the merits of this approach is provided.}, Address = {Estes Park, CO}, Author = {De Jong, Kenneth A. and William M. Spears and Diana F. Gordon}, Booktitle = {Proceedings of FOGA94}, Download1 = {PDF,http://www.cs.uwyo.edu/~wspears/papers/foga94/foga94.pdf}, Pages = {115--137}, Publisher = {Morgan Kaufmann}, Title = {Using Markov Chains to Analyze GAFOs}, Year = 1994 } @article{de-jong.spears:a-formal1992, Abstract = {On the basis of early theoretical and empirical studies, genetic algorithms have typically used 1 and 2-point crossover operators as the standard mechanisms for implementing recombination. However, there have been a number of recent studies, primarily empirical in nature, which have shown the benefits of crossover operators involving a higher number of crossover points. From a traditional theoretical point of view, the most surprising of these new results relate to uniform crossover, which involves on the average L / 2 crossover points for strings of length L. In this paper we extend the existing theoretical results in an attempt to provide a broader explanatory and predictive theory of the role of multi-point crossover in genetic algorithms. In particular, we extend the traditional disruption analysis to include two general forms of multi-point crossover: n-point crossover and uniform crossover. We also analyze two other aspects of multi-point crossover operators, namely, their recombination potential and exploratory power. The results of this analysis provide a much clearer view of the role of multi-point crossover in genetic algorithms. The implications of these results on implementation issues and performance are discussed, and several directions for further research are suggested.}, Author = {De Jong, Kenneth A. and William M. Spears}, Download1 = {PDF,http://www.cs.uwyo.edu/~wspears/papers/annals92.pdf}, Journal = {Annals of Mathematics and Artificial Intelligence Journal}, Pages = {1--26}, Title = {A Formal Analysis of the Role of Multi-Point Crossover in Genetic Algorithms}, Year = 1992 } @inproceedings{de-jong.spears:an-analysis1990, Abstract = {In this paper we present some theoretical and empirical results on the interacting roles of population size and crossover in genetic algorithms. We summarize recent theoretical results on the disruptive effect of two forms of multi-point crossover: npoint crossover and uniform crossover. We then show empirically that disruption analysis alone is not sufficient for selecting appropriate forms of crossover. However, by taking into account the interacting effects of population size and crossover, a general picture begins to emerge. The implications of these results on implementation issues and performance are discussed, and several directions for further research are suggested.}, Address = {University of Dortmund}, Author = {De Jong, Kenneth A. and William M. Spears}, Booktitle = {International Workshop Parallel Problem Solving from Nature}, Download1 = {PDF,http://www.cs.uwyo.edu/~wspears/papers/ppsn90.pdf}, Pages = {38--47}, Title = {An Analysis of the Interacting Roles of Population Size and Crossover in Genetic Algorithms}, Year = 1990 } @inproceedings{de-jong.spears:learning1991, Abstract = {In this paper we explore the use of an adaptive search technique (genetic algorithms) to construct a system GABIL which continually learns and refines concept classification rules from its interaction with the environment. The performance of the system is measured on a set of concept learning problems and compared with the performance of two existing systems: ID5R and C4.5. Preliminary results support that, despite minimal system bias, GABIL is an effective concept learner and is quite competitive with ID5R and C4.5 as the target concept increases in complexity.}, Address = {Sidney, Australia}, Author = {De Jong, Kenneth A. and William M. Spears}, Booktitle = {Proceedings of the International Joint Conference on Artificial Intelligence}, Download1 = {PDF,http://www.cs.uwyo.edu/~wspears/papers/ijcai91.pdf}, Pages = {651--656}, Title = {Learning Concept Classification Rules Using Genetic Algorithms}, Year = 1991 } @inproceedings{de-jong.spears:on-the-state1993, Abstract = {In the past few years the evolutionary computation landscape has been rapidly changing as a result of increased levels of interaction between various research groups and the injection of new ideas which challenge old tenets. The effect has been simultaneously exciting, invigorating, annoying, and bewildering to the old-timers as well as the new-comers to the field. Emerging out of all of this activity are the beginnings of some structure, some common themes, and some agreement on important open issues. We attempt to summarize these emergent properties in this paper.}, Address = {Urbana-Champaign, IL}, Author = {De Jong, Kenneth A. and William M. Spears}, Booktitle = {Proceedings of the 1993 International Conference on Genetic Algorithms}, Download1 = {PDF,http://www.cs.uwyo.edu/~wspears/papers/icga93.pdf}, Pages = {618--623}, Title = {On the State of Evolutionary Computation}, Year = 1993 } @inproceedings{de-jong.spears:using1989, Abstract = {A strategy for using Genetic Algorithms (GAs) to solve NP-complete problems is presented. The key aspect of the approach taken is to exploit the observation that, although all NP-complete problems are equally difficult in a general computational sense, some have much better GA representations than others, leading to much more successful use of GAs on some NP-complete problems than on others. Since any NP-complete problem can be mapped into any other one in polynomial time, the strategy described here consists of identifying a canonical NP-complete problem on which GAs work well, and solving other NP-complete problems indirectly by mapping them onto the canonical problem. Initial empirical results are presented which support the claim that the Boolean Satisfiability Problem (SAT) is a GAeffective canonical problem, and that other NPcomplete problems with poor GA representations can be solved efficiently by mapping them first onto SAT problems.}, Address = {George Mason University, Fairfax, VA}, Author = {De Jong, Kenneth A. and William M. Spears}, Booktitle = {Proceedings of the Third International Conference on Genetic Algorithms}, Download1 = {PDF,http://www.cs.uwyo.edu/~wspears/papers/icga89.pdf}, Pages = {124--132}, Title = {Using Genetic Algorithms to Solve NP-Complete Problems}, Year = 1989 } @incollection{deb.spears:speciation1997, Author = {K. Deb and William M. Spears}, Booktitle = {The Handbook of Evolutionary Computation}, Download1 = {PostScript,http://www.aic.nrl.navy.mil/~spears/papers/ece1_3.ps}, Download2 = {GZipped PostScript,http://www.aic.nrl.navy.mil/~spears/papers/ece1_3.ps.gz}, Editor = {T. Baeck, D. Fogel and Z. Michalewicz}, Note = {A portion of this paper entitled ``Speciation Using Tag Bits'' is available online}, Publisher = {IOP Publishing and Oxford University Press}, Title = {Speciation Methods}, Year = 1997 } @inproceedings{grajdeanu.jong:fixed2003, Abstract = {A well-known problem in dealing with noisy tness landscapes is the diculty in deciding a priori how many samples per individual to take in order to get a useful estimate of an individual's tness. This is particularly im- portant when an EA is given a priori a xed total budget of samples and must trade o increased accuracy of tness estimates over fewer generations vs. less accurate tness estimates over more generations. This pa- per investigates a technique for dynamically adapting the tness estimation accuracy as evolution proceeds and compares the results of such adaptive runs with others in which the accuracy is chosen a priori and kept con- stant.}, Author = {Adrian Grajdeanu and Kenneth A. De Jong}, Booktitle = {Genetic and Evolutionary Computation Conference (LBP)}, Download1 = {PDF,http://www.cs.gmu.edu/~eclab/papers/grajdeanu03fixed.pdf}, Title = {Fixed Budget Allocation Strategies for Noisy Fitness Landscapes}, Year = 2003 } @inproceedings{grajdeanu.jong:improving2004, Abstract = {Choosing representations and operators that preserve locality between genotype and phenotype space is an important goal in EA design. In the GA literature there has been considerable discussion of this issue with respect to the choice between standard binary encoding and Gray codes. In this paper we argue that an important and unappreciated aspect of such discussions is the degree to which locality preservation is isotropic in phenotype space (i.e., independent of location in phenospace). We show that using a traditional bit-flip mutation operator with either of these two representations results in rather weak isotropic locality. These insights lead to the design of a new binary mutation operator that increases isotropic locality. The results from an initial set of experiments supports the hypothesis that this improvement in isotropic locality leads to improvements in GA performance as well.}, Author = {Adrian Grajdeanu and Kenneth A. De Jong}, Booktitle = {Genetic and Evolutionary Computation Conference}, Download1 = {PDF,http://www.cs.gmu.edu/~eclab/papers/grajdeanu04improving.pdf}, Editor = {K. Deb et al.}, Pages = {1186--1196}, Publisher = {Springer-Verlag}, Title = {Improving the Locality Properties of Binary Representations}, Volume = {LNCS 3102}, Year = 2004 } @techreport{grajdeanu.kumar:a-novel2006, Abstract = {A novel developmental system designed to facilitate the study of generative encoding-based evolutionary design is presented. An aim of this work is to explore the simplest system and genetic encoding capable of exhibiting phenomena akin to size regulation and self-repair in developmental biology. Here we show that in the absence of complicated developmental mechanisms - such as ligand-receptor mediated cell signaling, asymmetric or mitotic spindle-based cell division, cell motility, or cellular forces - a simple rule-like GRN encoding and a simple internal/external chemical exchange mechanism confer the ability to generate patterns that are size regulated, and able to self-repair when damaged. In addition, we analyze the mechanisms that achieve size regulation and self-repair.}, Author = {Adrian Grajdeanu and Sanjeev Kumar}, Download1 = {PDF,http://www.cs.gmu.edu/~eclab/papers/grajdeanu06novel.pdf}, Institution = {Developmental Systems; AAAI Fall Symposium}, Number = {FS-06-03}, Pages = {24--30}, Title = {A Novel Developmental System for the Study of Evolutionary Design}, Year = 2006 } @techreport{grajdeanu:modeling2007, Abstract = {The present report details a model for implementing diffusion in a discrete space. Formulated in 2004 in support of artificial development experiments, the model is mathematically justified starting from the diffusion equation. It has physical plausibility and handles well different shaped and sized diffusion neighborhoods in the sense that it achieves isotropic diffusion in spite of the bias introduced by the discretizing grid. It provides one formulation that encapsulates the diffusion neighborhood details and renders it applicable to linear, planar, spatial and even n-dimensional constructs.}, Author = {Adrian Grajdeanu}, Institution = {George Mason University}, Number = {GMU-CS-TR-2008-1}, Title = {Modeling Diffusion in a Discrete Environment}, Year = 2007, Download1 = {PDF, http://cs.gmu.edu/~tr-admin/papers/GMU-CS-TR-2007-1.pdf} } @article{grefenstette.ramsey.ea:learning1990, Author = {John J. Grefenstette and Connie L. Ramsey and Alan C. Schultz}, Download1 = {Zipped PostScript,ftp://ftp.aic.nrl.navy.mil/pub/papers/1990/AIC-90-010.ps.Z}, Journal = {Machine Learning}, Pages = {355--381}, Publisher = {Kluwer Academic Publishers}, Title = {Learning sequential decision rules using simulation models and competition}, Year = 1990 } @inproceedings{grefenstette.schultz:evolutionary1994, Address = {New Brunswick, NJ}, Author = {John Grefenstette and Alan Schultz}, Booktitle = {Machine Learning Workshop on Robot Learning}, Download1 = {GZipped PostScript,ftp://ftp.aic.nrl.navy.mil/pub/papers/1994/AIC-94-014.ps.gz}, Title = {Evolutionary Approach to Learning in Robots}, Year = 1994 } @inproceedings{imam.vafaie:an-empirical1994, Author = {I. Imam and Haleh Vafaie}, Booktitle = {Proceedings of the Florida AI Research Symposium (FLAIRS-94)}, Download1 = {PDF,papers/flairs94.pdf}, Download2 = {PostScript,papers/flairs94.ps}, Download3 = {GZipped PostScript,papers/flairs94.ps.gz}, Pages = {66--70}, Title = {An Empirical Comparison Between Global And Greedy-like Search For Feature Selection}, Year = 1994 } @inproceedings{jansen.wiegand:bridging2004, Abstract = {While the gap between theory and practice is slowly closing, the evolutionary computation community needs to concentrate more heavily on the middle ground. This paper defends the position that contemporary analytical tools facilitate such a concentration. Empirical research can be improved by considering modern analytical techniques in experimental design. In addition, formal analytical extensions of empirical works are possible. We justify our position by way of a constructive example: we consider a recent empirically-based research paper and extend it using modern techniques of asymptotic analysis of run time performance of the algorithms and problems investigated in that paper. The result is a more general understanding of the performance of these algorithms for any size of input, as well as a better understanding of the underlying reasons for some of the previous results. Moreover, our example points out how important it is that empirical researchers motivate their parameter choices more clearly. We believe that providing theorists with empirical studies that are well-suited for formal analysis will help bridge the gap between theory and practice, benefitting the empiricist, the theorist, and the community at large.}, Author = {Thomas Jansen and R. Paul Wiegand}, Booktitle = {Parallel Problem Solving from Nature -- PPSN-2004}, Pages = {61--71}, Publisher = {Springer}, Title = {Bridging the Gap Between Theory and Practice}, Year = 2004 } @inproceedings{jansen.wiegand:exploring2003, Abstract = {Using a well-known cooperative coevolutionary function optimization framework, a very simple cooperative coevolutionary (1+1) EA is defined. This algorithm is investigated under the perspective of the expected optimization time. The focus is on the impact the cooperative coevolutionary approach has and on the possible advantage it may have over more traditional evolutionary approaches. Therefore, a systematic comparison between the expected optimization times of this coevolutionary algorithm and the ordinary (1+1) EA is presented. The main result is that separability of the objective function alone is is not sufficient to make the cooperative coevolutionary approach beneficial. By presenting a clear structured example function and analyzing the algorithms' performance, it is shown that the cooperative coevolutionary approach comes with new explorative possibilities. This can lead to an immense speed-up of the optimization.}, Author = {Jansen, Thomas and Wiegand, R. Paul}, Booktitle = {Proceedings of the 2003 Genetic and Evolutionary Computation Conference}, Download1 = {PDF,http://www.tesseract.org/paul/papers/gecco03-ccea.pdf}, Download2 = {PostScript,http://www.tesseract.org/paul/papers/gecco03-ccea.ps}, Publisher = {Springer}, Title = {Exploring the Explorative Advantage of the CC (1+1) EA}, Year = 2003 } @article{jansen.wiegand:the-cooperative2004, Abstract = {Coevolutionary algorithms are variants of traditional evolutionary algorithms and are often considered more suitable for certain kinds of complex tasks than non-coevolutionary methods. One example is a general cooperative coevolutionary framework for function optimization. This paper presents a thorough and rigorous introductory analysis of the optimization potential of cooperative coevolution. Using the cooperative coevolutionary framework as a starting point, the CC (1+1) EA is defined and investigated from the perspective of the expected optimization time. The research concentrates on separability, a key property of objective functions. We show that separability alone is not sufficient to yield any advantage of the CC (1+1) EA over its traditional, non-coevolutionary counterpart. Such an advantage is demonstrated to have its basis in the increased explorative possibilities of the cooperative coevolutionary algorithm. For inseparable functions, the cooperative coevolutionary set-up can be harmful. We prove that for some objective functions the CC (1+1) EA fails to locate a global optimum with overwhelming probability, even in infinite time; however, inseparability alone is not sufficient for an objective function to cause difficulties. It is demonstrated that the CC (1+1) EA may perform equal to its traditional counterpart, and may even outperform it on certain inseparable functions.}, Author = {Jansen, Thomas and Wiegand, R. Paul}, Journal = {Evolutionary Computation}, Keywords = {cooperative coevolution, run time analysis, separability, global optimization}, Number = 4, Publisher = {MIT Press}, Title = {The Cooperative Coevolutionary (1+1) EA}, Volume = 12, Pages = {405--434}, Year = 2004 } @inproceedings{jong.arciszewski.ea:an-overview1999, Author = {Kenneth De Jong and Tomasz Arciszewski and H. Vyas}, Booktitle = {Proceedings of 6th Workshop of the European Group for Structural Engineering Applications of AI}, Title = {An Overview of Evolutionary Computation and Its Applications to Engineering Design}, Year = 1999 } @book{jong:evolutionary2006, Abstract = {Evolutionary computation, the use of evolutionary systems as computational processes for solving complex problems, is a tool used by computer scientists and engineers who want to harness the power of evolution to build useful new artifacts, by biologists interested in developing and testing better models of natural evolutionary systems, and by artificial life scientists for designing and implementing new artificial evolutionary worlds. In this clear and comprehensive introduction to the field, Kenneth De Jong presents an integrated view of the state of the art in evolutionary computation. Although other books have described such particular areas of the field as genetic algorithms, genetic programming, evolution strategies, and evolutionary programming, Evolutionary Computation is noteworthy for considering these systems as specific instances of a more general class of evolutionary algorithms. This useful overview of a fragmented field is suitable for classroom use or as a reference for computer scientists and engineers.}, Author = {Kenneth A. De Jong}, Publisher = {MIT Press}, Title = {Evolutionary Computation: A Unified Approach}, Year = 2006 } @inproceedings{keller.antonisse:prediction-based1987, Address = {Washington, DC}, Author = {K. S. Keller and Jim Antonisse}, Booktitle = {Proceedings of the 3rd Annual Expert Systems in Government Conference}, Title = {Prediction-based Competitive Learning in the M2 System}, Year = 1987 } @inproceedings{kennedy.spears:matching1998, Author = {James Kennedy and William M. Spears}, Booktitle = {IEEE International Conference on Evolutionary Computation}, Download1 = {PostScript,http://www.aic.nrl.navy.mil/~spears/papers/wcci98.ps}, Download2 = {GZipped PostScript,http://www.aic.nrl.navy.mil/~spears/papers/wcci98.ps.gz}, Pages = {78--83}, Title = {Matching Algorithms to Problems: An Experimental Test of the Particle Swarm and Some Genetic Algorithms on the Multimodal Problem Generator}, Year = 1998 } @inproceedings{kicinger.arciszewski.ea:conceptual2003, Abstract = {This paper describes a new design paradigm, evolutionary structural design, that involves the entire design process, including conceptual and detailed design stages. In this paper, first a brief overview of the fundamentals of evolutionary computation is provided. Next, the concept of evolutionary structural design and its principia are discussed. Inventor 2001 is described in the following section. It is an experimental research and design system based on evolutionary computation. The system has been developed by the authors at George Mason University for applications in the design of tall buildings. Inventor 2001 allows for the conducting of evolutionary structural design, including the generation of structural concepts and the detailed design, analysis of internal forces, dimensioning, and optimization. Selected specific research results are also provided, including a discussion of the discovered emergent structural shaping patterns that are surprisingly consistent with the state of the art in structural shaping of steel skeleton structures of tall buildings. Finally, the initial research conclusions are provided.}, Address = {Singapore}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Booktitle = {Proceedings of the 2nd International Specialty Conference on the Conceptual Approach to Structural Design, Milan, Italy, July 1-2, 2003}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerCASD2003.pdf}, Keywords = {conceptual design and evolutionary design and tall buildings and evolutionary computation and structural design}, Pages = {529-536}, Publisher = {CI-Premier PTE Ltd.}, Title = {Conceptual design in structural engineering: an evolutionary computation approach}, Volume = 2, Year = 2003 } @inproceedings{kicinger.arciszewski.ea:distributed2004, Abstract = {This paper presents results of a study on distributed, or parallel, evolutionary computation in the topological design of steel structural systems in tall buildings. It describes results of extensive experimental research on various parallel evolutionary architectures applied to a complex structural design problem. The experiments were conducted using Inventor 2003, a network-based evolutionary design support tool developed at George Mason University. First, a general introduction to evolutionary computation is provided with an emphasis on recent developments in parallel evolutionary architectures. Next, a discussion of conceptual design of steel structural systems in tall buildings is presented. Further, Inventor 2003 is briefly introduced as well as its design representation and evolutionary computation characteristics. Next, the results obtained from systematic design experiments conducted with Inventor 2003 are discussed. The objective of these experiments was to qualitatively and quantitatively investigate evolution of steel structural systems in tall buildings during a distributed evolutionary design process as well as to compare efficiency and effectiveness of various parallel evolutionary architectures with the traditional evolutionary design approaches. Two connectivity topologies (ring topology and fully-connected topology) have been investigated for four populations of structural designs evolving in parallel and using various migration strategies. Also, results of the initial sensitivity studies are reported in which two ways of initializing distributed evolutionary design processes were investigated, using either arbitrarily selected designs as initial parents or randomly generated ones. Finally, initial research conclusions are presented.}, Address = {Weimar, Germany}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Booktitle = {Proceedings of the Xth International Conference on Computing in Civil and Building Engineering (ICCCBE-X), Weimar, Germany, June 2-4, 2004}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerICCCBE-X2004.pdf}, Editor = {Beucke, Karl and Firmenich, Berthold and Donath, Dirk and Fruchter, Renate and Roddis, Kim}, Isbn = {386068213X}, Keywords = {evolutionary computation and evolutionary design and distributed evolutionary algorithm and island-model and tall buildings}, Pages = 190, Publisher = {VDG}, Title = {Distributed evolutionary design: island-model based optimization of steel skeleton structures in tall buildings}, Year = 2004 } @inproceedings{kicinger.arciszewski.ea:emergent2004, Abstract = {The paper introduces an integrated research and design support tool, called Emergent Designer, developed at George Mason University. It is a tool that implements models of various complex systems, including cellular automata and evolutionary algorithms, to represent engineering systems and design processes. The system is intended for conducting design experiments in the area of structural design and for the analysis of their results. It implements state-of-the-art representations supporting generation of novel design concepts and efficient mechanisms for their subsequent optimization at the topology and sizing level. It also implements advanced methods, models, and tools from statistics and from the linear as well as nonlinear time series analysis to conduct the analysis of the design processes. Thus, it is a versatile tool that can be used both as a state-of-the-art design support tool and as an advanced research tool equipped with the methods and tools for the analysis of the design processes and of the obtained experimental results.}, Address = {Cambridge, MA, USA}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Booktitle = {Proceedings of the Workshop on the Implementation Issues in Generative Design Systems at the First International Conference on Design Computing and Cognition (DCC'04), Massachusetts Institute of Technology, Cambridge, MA, July 17-21, 2004}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerDCC2004.pdf}, Editor = {Caldas, Luisa G. and Duarte, Jose P.}, Keywords = {evolutionary computation and generative representations and morphogenic evolutionary design and cellular automata and design support tools and evolutionary design}, Pages = {93-112}, Title = {Emergent Designer: An integrated research and design support tool based on models of complex systems}, Year = 2004 } @inproceedings{kicinger.arciszewski.ea:emergent2004, Abstract = {The paper introduces an integrated research and design support tool, called Emergent Designer, developed at George Mason University. It is a tool that implements models of various complex systems, including cellular automata and evolutionary algorithms, to represent engineering systems and design processes. The system is intended for conducting design experiments in the area of structural design and for the analysis of their results. It implements state-of-the-art representations supporting generation of novel design concepts and efficient mechanisms for their subsequent optimization at the topology and sizing level. It also implements advanced methods, models, and tools from statistics and from the linear as well as nonlinear time series analysis to conduct the analysis of the design processes. Thus, it is a versatile tool that can be used both as a state-of-the-art design support tool and as an advanced research tool equipped with the methods and tools for the analysis of the design processes and of the obtained experimental results.}, Author = {Rafal Kicinger and Tomasz Arciszewski and Kenneth De Jong}, Booktitle = {Proceedings of the Workshop on the Implementation Issues in Generative Design Systems at the First International Conference on Design Computing and Cognition}, Editor = {L. G. Caldis and J. P. Duarte}, Pages = {93 -- 112}, Publisher = {MIT Press}, Title = {Emergent Designer: Generative Design in Structural Engineering}, Year = 2004 } @article{kicinger.arciszewski.ea:evolutionary2005, Abstract = {Evolutionary computation is emerging as a new engineering computational paradigm, which may significantly change the present structural design practice. For this reason, an extensive study of evolutionary computation in the context of structural design has been conducted in the Information Technology and Engineering School at George Mason University and its results are reported here. First, a general introduction to evolutionary computation is presented and recent developments in this field are briefly described. Next, the field of evolutionary design is introduced and its relevance to structural design is explained. Further, the issue of creativity/novelty is discussed and possible ways of achieving it during a structural design process are suggested. Current research progress in building engineering systems' representations, one of the key issues in evolutionary design, is subsequently discussed. Next, recent developments in constraint-handling methods in evolutionary optimization are reported. Further, the rapidly growing field of evolutionary multiobjective optimization is presented and briefly described. An emerging subfield of coevolutionary design is subsequently introduced and its current advancements reported. Next, a comprehensive review of the applications of evolutionary computation in structural design is provided and chronologically classified. Finally, a summary of the current research status and a discussion on the most promising paths of future research are also presented.}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerCAS2005.pdf}, Issn = 00457949, Journal = {Computers & Structures}, Keywords = {evolutionary computation and structural design and review and engineering design and creative design and inventive design and design representations and constraint-handling methods and multiobjective optimization and coevolutionary design}, Number = {23-24}, Pages = {1943-1978}, Title = {Evolutionary computation and structural design: A survey of the state of the art}, Volume = 83, Year = 2005 } @article{kicinger.arciszewski.ea:evolutionary2005, Abstract = {This paper presents results of a study on evolutionary computation in the design of the steel structural systems of tall buildings. It describes results of extensive research on both short-term (up to a few hundred generations) and long-term evolutionary design processes (at least a few thousand generations). The experiments were conducted with Inventor 2001, an evolutionary design support tool developed at George Mason University, for generating conceptual and detailed designs of steel structural systems in tall buildings. First, the paper discusses conceptual design of steel structural systems in tall buildings and briefly introduces Inventor 2001 as well as its design representation and evolutionary computation characteristics. Next, it provides the results obtained from systematic parametric design experiments conducted with Inventor 2001. The objective of these experiments was to qualitatively and quantitatively investigate evolution of steel structural systems of tall buildings during a multistage evolutionary design process as well as the influence of various evolutionary computation parameters. Mutation and crossover rates, population size, the length of the evolutionary processes, and the importance of symmetry requirement have been analyzed and results produced. Emergence of structural shaping patterns has been also studied and several interesting patterns found in the evolutionary design process. Finally, research conclusions are presented as well as recommendations for further research and development of evolutionary design support tools.}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Journal = {Journal of Computing in Civil Engineering}, Keywords = {evolutionary computation and optimization and evolutionary design and tall buildings and structural design}, Number = 3, Pages = {223-238}, Title = {Evolutionary design of steel structures in tall buildings}, Volume = 19, Year = 2005 } @inproceedings{kicinger.arciszewski.ea:generative2005, Abstract = {This paper proposes a new approach to representing structural system inspired by various models of complex systems. Several types of generative representations of steel structural systems are provided and empirically investigated. These representations utilize various kinds of cellular automata to generate design concepts of steel structures in tall buildings. In the paper, a brief overview of the state-of-the-art in cellular automata and generative design is presented. Next, several types of generative representations of steel structural systems in tall buildings are described. The paper also reports the results of several design experiments. They have shown that generative representations produce novel structural shaping patterns which are qualitatively different than the patterns obtained using traditionally used parameterized representations. They also significantly improve the performance of evolutionary algorithms optimizing the structural systems. Finally, research conclusions are presented and most promising paths of future research are discussed.}, Address = {Reston, VA}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Booktitle = {Proceedings of the 2005 ASCE International Conference on Computing in Civil Engineering}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerASCE2005.pdf}, Editor = {Soibelman, Lucio and Pena-Mora, Feniosky}, Isbn = 0784407940, Keywords = {generative representations and morphogenic evolutionary design and evolutionary computation and structural design and cellular automata and morphogenesis}, Month = {July}, Publisher = {American Society of Civil Engineers Press}, Title = {Generative design in structural engineering}, Year = 2005 } @inproceedings{kicinger.arciszewski.ea:morphogenesis2004, Abstract = {This paper provides the initial results of a study on the applications of generative cellular automata-based representations in evolutionary structural design. First, recent developments in evolutionary design representations and an overview of cellular automata are presented. Next, a complex problem of topological design of steel structural systems in tall buildings is briefly described. Further, morphogenic evolutionary design is introduced and exemplified by cellular automata representations. The paper also reports the initial results of several structural design experiments whose objective was to determine feasibility of the proposed approach. Finally, initial research conclusions are provided.}, Address = {Piscataway, NJ}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Booktitle = {Proceedings of the Congress on Evolutionary Computation (CEC'2004), Portland, Oregon, June 19-23, 2004}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerCEC2004.pdf}, Isbn = 0780385152, Keywords = {evolutionary computation and morphogenesis and morphogenic evolutionary design and evolutionary design and cellular automata and tall buildings and structural design}, Pages = {411-418}, Publisher = {IEEE Press}, Title = {Morphogenesis and structural design: cellular automata representations of steel structures in tall buildings}, Year = 2004 } @incollection{kicinger.arciszewski.ea:morphogenic2004, Abstract = {This paper provides the initial results of a study on the applications of cellular automata representations in evolutionary design of topologies of steel structural systems in tall buildings. In the paper, a brief overview of the state of the art in cellular automata and evolutionary design representations is presented. Next, morphogenic evolutionary design is introduced and illustrated by several types of cellular automata representations. Further, Emergent Designer, a unique evolutionary design tool developed at George Mason University, is briefly described. It is an integrated research and design support tool which applies models of complex adaptive systems to represent engineering systems and analyze design processes. The paper also reports the initial results of several structural design experiments conducted with Emergent Designer. The objective of the experiments was to determine feasibility of various types of cellular automata representations in topological structural optimization. Finally, initial research conclusions and recommendations for the further research are provided.}, Address = {London, UK}, Author = {Kicinger, Rafal and Arciszewski, Tomasz and De Jong, Kenneth A.}, Booktitle = {Adaptive Computing in Design and Manufacture VI}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerACDM2004.pdf}, Editor = {Parmee, Ian C.}, Isbn = 1852338296, Keywords = {cellular automata and morphogenesis and morphogenic evolutionary design and evolutionary computation and evolutionary design and tall buildings and structural design}, Pages = {25-38}, Publisher = {Springer-Verlag}, Title = {Morphogenic evolutionary design: cellular automata representations in topological structural design}, Year = 2004 } @inproceedings{kicinger.arciszewski.ea:parameterized2005, Abstract = {Any computational approach to design, including the use of evolutionary algorithms, requires the transformation of the domain-specific knowledge into a formal design representation. This is a difficult and still not completely understood process. Its critical part is the choice of a type of design representation. The paper addresses this important issue by presenting and discussing results of a large number of design experiments in which parameterized and generative representations were used. Particularly, their computational and design related advantages and disadvantages were investigated and compared. Evolutionary design experiments reported in this paper considered two classes of structural design problems, including the design of a wind bracing system and the design of an entire structural system in a tall building. Parameterized and generative representations of the structural systems were introduced and their basic features discussed. The generative representations investigated in the paper were inspired by the processes of morphogenesis occurring in nature. Specifically, one-dimensional cellular automata were used to develop, or "grow," structural designs from the corresponding "design embryos." The conducted research led to three major conclusions. First, generative representations based on cellular automata proved to scale well with the size of the considered design problems. Second, generative representations outperformed parameterized representations in minimizing weight of the structural systems in our problem domain by generating better designs and finding them faster. Finally, extensive experimental studies showed significant differences in optimal settings for evolutionary design experiments for the two representation types. The rate of mutation operator, the size of the parent population, and the type of the evolutionary algorithm were identified as the evolutionary parameters having the largest impact on the performance of evolutionary design processes in our problem domain.}, Author = {Rafal Kicinger and Tomasz Arciszewski and Kenneth De Jong}, Booktitle = {Proceedings of Genetic and Evolutionary Computation Conference -- GECCO-2005}, Download1 = {PDF,http://cs.gmu.edu/~eclab/papers/kicinger05parameterized.pdf}, Pages = {2007--2014}, Publisher = {ACM Press}, Title = {Parameterized versus Generative Representations in Structural Design: An Empirical Comparison}, Year = 2005 } @inproceedings{kicinger.arciszewski:multiobjective2004, Abstract = {This paper presents initial results of a study on the application of evolutionary multi-objective optimization methods in the design of the steel structural systems of tall buildings. In the paper, a brief overview of the state-of-the-art in evolutionary multi-objective optimization in structural engineering is provided. Next, conceptual design of steel structural systems in tall buildings is overviewed and the representations of steel structural systems used in the paper are discussed. Furthermore, Emergent Designer, a unique evolutionary design tool developed at George Mason University, is briefly described. It is an integrated research and design support tool which applies models of complex adaptive systems to represent engineering systems and to analyze design processes and their results. The paper also presents the results of several multi-objective structural design experiments conducted with Emergent Designer in which steel structural systems in tall buildings were optimized with respect to their total weight and maximum deflection (two-objective minimization problem). The goal of these experiments was to determine feasibility of evolutionary multi-objective optimization of steel structural systems of tall buildings as well as to qualitatively and quantitatively compare the results with the previous findings obtained with single-objective evolutionary optimization methods. Finally, initial research conclusions are presented as well as promising research directions.}, Address = {Reston, VA}, Author = {Kicinger, Rafal and Arciszewski, Tomasz}, Booktitle = {Proceedings of the AIAA 1st Intelligent Systems Technical Conference, Chicago, IL, September 20-23, 2004}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerAIAA-IS2004.pdf}, Keywords = {evolutionary computation and multiobjective optimization and evolutionary design and tall buildings}, Pages = {AIAA 2004-6438}, Publisher = {American Institute of Aeronautics and Astronautics Press}, Title = {Multiobjective evolutionary design of steel structures in tall buildings}, Year = 2004 } @inproceedings{kicinger.de-jong.ea:long2002, Address = {Duesseldorf, Germany}, Author = {Kicinger, Rafal and De Jong, Kenneth A. and Arciszewski, Tomasz}, Booktitle = {Advances in Intelligent Computing in Engineering. Proceedings of the 9th International Workshop of the European Group for Intelligent Computing in Engineering, Darmstadt, Germany, August 1-2, 2002}, Download1 = {PDF,http://www.kicinger.com/publications/pdf/KicingerEG-ICE2002.pdf}, Editor = {Schnellenbach-Held, Martina and Denk, Heiko}, Isbn = 3183180049, Keywords = {evolutionary computation and evolutionary design and structural design and tall buildings and optimization}, Pages = {184-195}, Publisher = {VDI Verlag}, Title = {Long term versus short term evolutionary design}, Year = 2002 } @phdthesis{kicinger:emergent2004, Abstract = {For a long time, engineering design research has been focused on the development of various design theories, methodologies, methods, tools, and procedures. The design methods have been subsequently used by engineers to more efficiently design artifacts. However, as the artifacts have grown in complexity, the need for new methods has become obvious. Also, in a nowadays world, increased competition and globalization require organizations to reexamine traditional product development strategies. Traditional methods focused exclusively on the numerical optimality of produced artifacts, or their manufacturing processes, are no longer adequate. Creativity and innovation of designed artifacts provide organizations not only with a competitive advantage but are, in fact, a matter of their survival. This dissertation addresses this problem by posing and answering the question: "How can one construct an effective method for designing engineering systems that would support development of novel/creative designs and their efficient optimization?" It proposes a new and conceptually coherent design method, called Emergent Engineering Design. The proposed design method is inspired by the fundamental processes occurring in nature, which has arguably created the most fascinating designs known to humankind. All major phases of Emergent Engineering Design are represented by complex systems, including cellular automata and evolutionary algorithms, which have been successfully used to model the processes governing the complex behavior occurring in nature. In order to facilitate the development of the proposed design method, Emergent Engineering Design was implemented in a computer system called Emergent Designer. It is an integrated research and design support tool which applies models of complex systems to represent engineering systems and analyze design processes. Emergent Designer was used to conduct the empirical validation of the proposed design method for two classes of conceptual design problems in structural engineering. The extensive design experiments reported in this dissertation have shown that Emergent Engineering Design not only generates novel design concepts exhibiting remarkable structural shaping patterns but it also efficiently optimizes them.}, Address = {Fairfax, VA}, Author = {Kicinger, Rafal}, Download1 = {PDF,http://www.kicinger.com/dissertation/KicingerDissertationEdited.pdf}, Download2 = {Downloading instructions,http://mason.gmu.edu/~rkicinge/publications.html#Dissertation}, Keywords = {engineering design and morphogenic evolutionary design and design methods and complex systems and evolutionary computation and evolutionary design and cellular automata and optimization and creative design}, School = {George Mason University}, Title = {Emergent Engineering Design: Design creativity and optimality inspired by nature}, Type = {Ph.D. Dissertation}, Year = 2004 } @inproceedings{kumar:a-developmental2005, Abstract = {The need to build modular, scalable, and complex technologycapable of adaptation, self-assembly, and self-repair hasfuelled renewed interest in using approaches inspired by developmentalbiology. To meet this need, a new eld, calledComputational Development (CD), has emerged. Its focus ison adapting processes and mechanisms from developmentalbiology so as to help us build scalable, complex technology.Due to the embryonic nature of the eld, however, researchinvestigating the potential of such approaches for di erentproblem domains is crucial to its success. In this paper, theplausibility of applying a developmental biology-inspired approachto the demanding problem domain of reactive robotcontrol is explored. Using developmental genetics as a sourceof inspiration, a model of genetic regulatory networks is usedin conjunction with a spatially distributed evolutionary algorithmto evolve real-time robot controllers for tasks suchas general purpose obstacle avoidance.}, Author = {Sanjeev Kumar}, Booktitle = {Proceedings of the Second Workshop On Self-Organization in Representations For Evolutionary Algorithms: Building complexity from simplicity, GECCO-2005}, Download1 = {PDF,http://cs.gmu.edu/~eclab/papers/kumar05developmental.pdf}, Publisher = {ACM Press}, Title = {A Developmental Genetics-Inspired Approach to Robot Control}, Year = 2005 } @inproceedings{luke.balan.ea:mason:2003, Abstract = {Agent-based modeling (ABM) has transformed social science research by allowingresearchers to replicate or generate the emergence of empirically complex socialphenomena from a set of relatively simple agent-based rules at the micro-level.Swarm, RePast, Ascape, and others currently provide simulation environments forABM social science research. After Swarm — arguably the first widely used ABMsimulator employed in the social sciences — subsequent simulators have sought toenhance available simulation tools and computational capabilities by providingadditional functionalities and formal modeling facilities. Here we present MASON(Multi-Agent Simulator Of Neighborhoods), following in a similar tradition that seeksto enhance the power and diversity of the available scientific toolkit in computationalsocial science. MASON is intended to provide a core of facilities useful not only tosocial science but to other agent-based modeling fields such as artificial intelligenceand robotics. We believe this can foster useful “cross-pollination” between suchdiverse disciplines, and further that MASON's additional facilities will becomeincreasing important as social complexity simulation matures and grows into newapproaches. We illustrate the new MASON simulation library with a replication ofHeatBugs and a demonstration of MASON applied to two challenging case studies:ant-like foragers and micro-aerial agents. Other applications are also beingdeveloped. The HeatBugs replication and the two new applications provide an idea ofMASON’s potential for computational social science and artificial societies.}, Author = {Sean Luke and Gabriel Catalin Balan and Liviu Panait and Claudio Cioffi-Revilla and Sean Paus}, Booktitle = {Proceedings of Agent 2003 Conference on Challenges in Social Simulation}, Title = {{MASON}: A {J}ava Multi-agent Simulation Library}, Year = 2003 } @inproceedings{luke.balan.ea:mason:2003, Address = {Atlanta, Georgia}, Author = {Sean Luke and Gabriel Catalin Balan and Liviu Panait}, Booktitle = {Proceedings of the Second International Workshop on the Mathematics and Algorithms of Social Insects}, Title = {{MASON}: A {J}ava Multi-agent Simulation Library}, Year = 2003 } @inproceedings{luke.balan.ea:population2003, Abstract = {With the exception of a small body of adaptive-parameter literature, evolutionary computation has traditionally favored keeping the population size constant through the course of the run. Unfortunately, genetic programming has an aging problem: for various reasons, late in the run the technique become less effective at optimization. Given a fixed number of evaluations, allocating many of them late in the run may thus not be a good strategy. In this paper we experiment with gradually decreasing the population size throughout a genetic programming run, in order to reallocate more evaluations to early generations. Our results show that over four problem domains and three different numbers of evaluations, decreasing the population size is always as good as, and frequently better than, various fixed-sized population strategies.}, Author = {Sean Luke and Gabriel Catalin Balan and Liviu Panait}, Booktitle = {Genetic and Evolutionary Computation -- GECCO-2003}, Editor = {E. Cant{\'u}-Paz and J. A. Foster and K. Deb and D. Davis and R. Roy and U.-M. O'Reilly and H.-G. Beyer and R. Standish and G. Kendall and S. Wilson and M. Harman and J. Wegener and D. Dasgupta and M. A. Potter and A. C. Schultz and K. Dowsland and N. Jonoska and J. Miller}, Pages = {1729--1739}, Publisher = {Springer}, Series = {LNCS}, Title = {Population Implosion in Genetic Programming.}, Volume = 2724, Year = 2003 } @inproceedings{luke.cioffi-revilla.ea:mason:2004, Abstract = {We introduce MASON, a fast, easily extendable, discreteevent multi-agent simulation toolkit in Java. MASON was designed to serve as the basis for a wide range of multiagent simulation tasks ranging from swarm robotics to machine learning to social complexity environments. MASON carefully delineates between model and visualization, allowing models to be dynamically detached from or attached to visualizers, and to change platforms mid-run. We describe the MASON system, its motivation, and its basic architectural design. We then discuss five applications of MASON we have built over the past year to suggest its breadth of utility.}, Author = {Sean Luke and Claudio Cioffi-Revilla and Liviu Panait and Keith Sullivan}, Booktitle = {Proceedings of the 2004 SwarmFest Workshop}, Download1 = {PDF,http://cs.gmu.edu/~eclab/projects/mason/publications/SwarmFest04.pdf}, Title = {MASON: A New Multi-Agent Simulation Toolkit}, Year = 2004 } @article{luke.hamahashi.ea:biology:1998, Abstract = {Computer science owes a huge debt to biological systems. The field itself came about largely as an attempt to understand and replicate the function and abilities of the brain, a complex biological creation. From this early lineage has sprung many subfields derived largely from biological metaphors: computer vision, neural networks, evolutionary computation, robotics, multi-agent studies, and much of artificial intelligence. In some areas, the computer has bested its biological counterparts in efficiency and simplicity. But for many domains, even after decades of hard work, the biological {"}real thing{"} is still superior to the artificial algorithms inspired by it.}, Author = {Sean Luke and Shugo Hamahashi and Koji Kyoda and Hiroki Ueda}, Download1 = {PDF,http://www.cs.gmu.edu/~sean/papers/biology.pdf}, Download2 = {GZipped PostScript,http://www.cs.gmu.edu/~sean/papers/biology.ps.gz}, Journal = {IEEE Intelligent Systems}, Keywords = {genetic algorithms, genetic programming, biological modelling, DNA}, Notes = {Invited Article. Argues for a revisitation of the biological roots behind artificial intelligence and evolutionary computation}, Size = {3 pages}, Title = {Biology: See It Again -- for the First Time}, Volume = 13, Number = 5, Pages = {6 -- 8}, Year = 1998 } @inproceedings{luke.hamahashi.ea:genetic1999, Abstract = {Much of evolutionary computation was inspired by Mendelian genetics. But modern genetics has since advanced considerably, revealing that genes are not simply parameter settings, but interactive cogs in a complex chemical machine. At the same time, an increasing number of evolutionary computation domains are evolving non-parameterized mechanisms such as neural networks or symbolic computer programs. As such, we think modern biological genetics offers much in helping us understand how to evolve such things. In this paper, we present a gene regulation model for Drosophila melanogaster. We then apply gene regulation to evolve deterministic finite-state automata, and show that our approach does well compared to past examples from the literature.}, Address = {Orlando, Florida, USA}, Author = {Sean Luke and Shugo Hamahashi and Hiroaki Kitano}, Booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference}, Download1 = {PDF,http://www.cs.gmu.edu/~sean/papers/gene-gecco99.pdf}, Download2 = {GZipped PostScript,http://www.cs.gmu.edu/~sean/papers/gene-gecco99.ps.gz}, Editor = {Wolfgang Banzhaf and Jason Daida and Agoston E. Eiben and Max H. Garzon and Vasant Honavar and Mark Jakiela and Robert E. Smith}, Isbn = {1-55860-611-4}, Keywords = {genetic programming and evolvable hardware}, Notes = {GECCO-99 A joint meeting of the eighth international conference on genetic algorithms (ICGA-99) and the fourth annual genetic programming conference (GP-99)}, Pages = {1098--1105}, Publisher = {Morgan Kaufmann}, Publisher_Address ={San Francisco, CA 94104, USA}, Title = {``Genetic'' Programming}, Year = 1999 } @inproceedings{luke.hohn.ea:co-evolving1997, Abstract = {Genetic Programming is a promising new method for automatically generating functions and algorithms through natural selection. In contrast to other learning methods, Genetic Programming's automatic programming makes it a natural approach for developing algorithmic robot behaviors. In this paper we present an overview of how we apply Genetic Programming to behavior-based team coordination in the RoboCup Soccer Server domain. The result is not just a hand-coded soccer algorithm, but a team of softbots which have learned on their own how to play a reasonable game of soccer.}, Address = {Nagoya, Japan}, Author = {Sean Luke and Charles Hohn and Jonathan Farris and Gary Jackson and James Hendler}, Booktitle = {Proceedings of the First International Workshop on RoboCup, at the International Joint Conference on Artificial Intelligence}, Download1 = {PDF,http://www.cs.gmu.edu/~sean/papers/robocupc.pdf}, Download2 = {GZipped PostScript,http://www.cs.gmu.edu/~sean/papers/robocupc.ps.gz}, Keywords = {genetic algorithms, genetic programming}, Size = {4 pages}, Title = {Co-evolving Soccer Softbot Team Coordination with Genetic Programming}, Year = 1997 } @article{luke.panait:a-comparison2006, Abstract = {Genetic programming has highlighted the problem of bloat, the uncontrolled growth of the average size of an individual in the population. The most common approach to dealing with bloat in tree-based genetic programming individuals is to limit their maximal allowed depth. An alternative to depth limiting is to punish individuals in some way based on excess size, and our experiments have shown that the combination of depth limiting with such a punitive method is generally more effective than either alone. Which such combinations are most effective at reducing bloat? In this article we augment depth limiting with nine bloat control methods and compare them with one another. These methods are chosen from past literature and from techniques of our own devising. Testing with four genetic programming problems, we identify where each bloat control method performs well on a per-problem basis, and under what settings various methods are effective independent of problem. We report on the results of these tests, and discover an unexpected winner in the cross-platform category.}, Author = {Sean Luke and Liviu Panait}, Journal = {Evolutionary Computation}, Number = 3, Pages = {309 -- 344}, Title = {A Comparison of Bloat Control Methods for Genetic Programming}, Volume = 14, Year = 2006 } @inproceedings{luke.panait:a-survey2001, Abstract = {This paper discusses and compares five major tree-generation algorithms for genetic programming, and their effects on fitness: RAMPED HALF-AND-HALF, PTC1, PTC2, RANDOM-BRANCH, and UNIFORM. The paper compares the performance of these algorithms on three genetic programming problems (11-Boolean Multiplexer, Artificial Ant, and Symbolic Regression), and discovers that the algorithms do not have a significant impact on fitness. Additional experimentation shows that tree size does have an important impact on fitness, and further that the ideal initial tree size is very different from that used in traditional GP.}, Author = {Luke, Sean and Panait, Liviu}, Booktitle = {Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) 2001}, Download1 = {PDF,http://www.cs.gmu.edu/~sean/papers/treegenalgs.pdf}, Download2 = {GZipped PostScript,http://www.cs.gmu.edu/~sean/papers/treegenalgs.ps.gz}, Pages = {81--88}, Publisher = {Morgan Kaufmann Publishers}, Title = {A survey and comparison of tree generation algorithms}, Year = 2001 } @inproceedings{luke.panait:alternative2004, Abstract = {Bloat control is an important aspect of evolutionary computation methods, such as genetic programming, which must deal with genomes of arbitrary size. We introduce three new methods for bloat control: Biased Multi-Objective Parsimony Pressure (BMOPP), the Waiting Room, and Death by Size. These methods are unusual approaches to bloat control, and are not only useful in various circumstances, but two of them suggest novel approaches to attack the problem. BMOPP is a more traditional parsimony-pressure style bloat control method, while the other two methods do not consider parsimony as part of the selection process at all, but instead penalize for parsimony at other stages in the evolutionary process. We find parameter settings for BMOPP and the Waiting Room which are effective across all tested problem domains. Death by Size does not appear to have this consistency, but we find it a useful tool as it has particular applicability to steady-state evolution.}, Author = {Sean Luke and Liviu Panait}, Booktitle = {Genetic and Evolutionary Computation Conference -- GECCO-2004}, Publisher = {Springer}, Title = {Alternative Bloat Control Methods}, Year = 2004 } @inproceedings{luke.panait:fighting2002, Abstract = {Many forms of parsimony pressure are parametric, that is final fitness is a parametric model of the actual size and raw fitness values. The problem with parametric techniques is that they are hard to tune to prevent size from dominating fitness late in the evolutionary run, or to compensate for problem-dependent nonlinearities in the raw fitness function. In this paper we briefly discuss existing bloat-control techniques, then introduce two new kinds of non-parametric parsimony pressure, Direct and Proportional Tournament. As their names suggest, these techniques are based on simple modifications of tournament selection to consider both size and fitness, but not together as a combined parametric equation. We compare the techniques against, and in combination with, the most popular genetic programming bloat-control technique, Koza-style depth limiting, and show that they are effective in limiting size while still maintaining good best-fitness-of-run results.}, Author = {Sean Luke and Liviu A. Panait}, Booktitle = {Parallel Problem Solving from Nature - PPSN VII (LNCS 2439)}, Editor = {Juan Julian Merelo Guervos et al}, Pages = {411--421}, Publisher = {Springer}, Title = {Fighting Bloat With Nonparametric Parsimony Pressure}, Year = 2002 } @inproceedings{luke.panait:is-the-perfect2002, Abstract = {Much of the genetic programming literature compares techniques using counts of ideal solutions found. These counts in turn form common comparison measures such as Koza's Computational Effort or cumulative Probability of Success. The use of these measures continues despite past warnings that they are not statistically valid. In this paper we too criticize the measures for serious statistical problems, and also argue that their motivational justification is faulty. We then present evidence suggesting that ideal solution counts are not necessarily positively related to best-fitness-of-run statistics: in fact they are often inversely correlated. Thus claims based on ideal solution counts can mislead readers into thinking techniques will provide superior final results, when in fact the opposite is true.}, Address = {New York}, Author = {Sean Luke and Liviu A. Panait}, Booktitle = {GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference}, Editor = {W. B. Langdon and E. Cant{\'u}-Paz and K. Mathias and R. Roy and D. Davis and R. Poli and K. Balakrishnan and V. Honavar and G. Rudolph and J. Wegener and L. Bull and M. A. Potter and A. C. Schultz and J. F. Miller and E. Burke and N. Jonoska}, Pages = {820--828}, Publisher = {Morgan Kaufmann Publishers}, Publisher_Address ={San Francisco, CA 94104, USA}, Title = {Is the Perfect the Enemy of the Good?}, Year = 2002 } @inproceedings{luke.panait:lexicographic2002, Abstract = {We introduce a technique called lexicographic parsimony pressure, for controlling the significant growth of genetic programming trees during the course of an evolutionary computation run. Lexicographic parsimony pressure modifies selection to prefer smaller trees only when fitnesses are equal (or equal in rank). This technique is simple to implement and is not affected by specific differences in fitness values, but only by their relative ranking. In two experiments we show that lexicographic parsimony pressure reduces tree size while maintaining good fitness values, particularly when coupled with Koza-style maximum tree depth limits.}, Address = {New York}, Author = {Sean Luke and Liviu A. Panait}, Booktitle = {GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference}, Editor = {W. B. Langdon and E. Cant{\'u}-Paz and K. Mathias and R. Roy and D. Davis and R. Poli and K. Balakrishnan and V. Honavar and G. Rudolph and J. Wegener and L. Bull and M. A. Potter and A. C. Schultz and J. F. Miller and E. Burke and N. Jonoska}, Pages = {829--836}, Publisher = {Morgan Kaufmann Publishers}, Publisher_Address ={San Francisco, CA 94104, USA}, Title = {Lexicographic Parsimony Pressure}, Year = 2002 } @inproceedings{luke.sharma.ea:finding2007, Abstract = {Model- and simulation-designers are often interested not in the optimum output of their system, but in understanding how the output is sensitive to different parameters. This can require an inefficient sweep of a multidimensional parameter space, with many samples tested in regions of the space where the output is essentially all the same, or a sparse sweep which misses crucial "interesting" regions where the output is strongly sensitive. In this paper we introduce a novel population-oriented approach to adaptive parameter sweeping which focuses its samples on these sensitive areas. The method is easy to implement and model-free, and does not require any previous knowledge about the space. In a weakened form the method can operate in non-metric spaces such as the space of genetic program trees. We demonstrate the method on three test problems, showing that it identifies regions of the space where the slope of the output is highest, and concentrates samples on those regions.}, Address = {London, England}, Author = {Sean Luke and Deepankar Sharma and Gabriel Catalin Balan}, Booktitle = {Proceedings of Genetic and Evolutionary Computation Conference}, Month = {July}, Title = {Finding Interesting Things}, Year = 2007 } @inproceedings{luke.spector:a-comparison1997, Abstract = {This paper presents a large and systematic body of data on the relative effectiveness of mutation, crossover, and combinations of mutation and crossover in genetic programming (GP). The literature of traditional genetic algorithms contains related studies, but mutation and crossover in GP differ from their traditional counterparts in significant ways. In this paper we present the results from a very large experimental data set, the equivalent of approximately 12,000 typical runs of a GP system, systematically exploring a range of parameter settings. The resulting data may be useful not only for practitioners seeking to optimize parameters for GP runs, but also for theorists exploring issues such as the role of {"}building blocks{"} in GP.}, Address = {Stanford University, CA, USA}, Author = {Sean Luke and Lee Spector}, Booktitle = {Genetic Programming 1997: Proceedings of the Second Annual Conference}, Download1 = {Figures,http://www.cs.gmu.edu/~sean/papers/comparison/figures1-2.ps.gz}, Download2 = {Figures2,http://www.cs.gmu.edu/~sean/papers/comparison/figures3-4.ps.gz}, Download3 = {PDF,http://www.cs.gmu.edu/~sean/papers/comparison/comparison.pdf}, Download4 = {GZipped PostScript,http://www.cs.gmu.edu/~sean/papers/comparison/comparison.ps.gz}, Editor = {John R. Koza and Kalyanmoy Deb and Marco Dorigo and David B. Fogel and Max Garzon and Hitoshi Iba and Rick L. Riolo}, Keywords = {Genetic Programming, Genetic Algorithms}, Month = {July}, Notes = {GP-97. 6-mux, lawn mower, symbolic regression, Santa Fe trail artificial ant. SEE ALSO luke:1998:rcxmGP. The Gzipped PostScript version (.ps.gz) does not come with figures; to get the figures for the PostScript version, use the figures URLs below}, Pages = {240--248}, Publisher = {Morgan Kaufmann}, Publisher_Address ={San Francisco, CA, USA}, Title = {A Comparison of Crossover and Mutation in Genetic Programming}, Year = 1997 } @inproceedings{luke.spector:a-revised1998, Abstract = {In [Luke and Spector 1997] we presented a comprehensive suite of data comparing GP crossover and point mutation over four domains and a wide range of parameter settings. Unfortunately, the results were marred by statistical flaws. This revision of the study eliminates these flaws, with three times as much the data as the original experiments had. Our results again show that crossover does have some advantage over mutation given the right parameter settings (primarily larger population sizes), though the difference between the two surprisingly small. Further, the results are complex, suggesting that the big picture is more complicated than is commonly believed.}, Address = {University of Wisconsin, Madison, Wisconsin, USA}, Author = {Sean Luke and Lee Spector}, Booktitle = {Genetic Programming 1998: Proceedings of the Third Annual Conference}, Download1 = {Figures,htt