Holland's Adaptation in Natural and Artificial Systems largely dealt with how systems, comprised of many self-interested entities, can and should adapt as a whole. This seminal book led to the last 25 years of work in geneticalgorithms (GAs) and related forms of evolutionary computation (EC). In recent years, the expansion of the Internet, other telecommunications technologies, and other large scale networks have led to a world where large numbers of semi-autonomous software entities (i.e., agents) will be interacting in an open, universal system. This development cast the importance of Holland's legacy in a new light. This paper argues that Holland's fundamental arguments, and the years of developments that have followed, have a direct impact on systems of general network agents, regardless of whether they explicitly exploit EC. However, it also argues that the techniques and theories of EC cannot be directly transferred to the world of general agents (rather than EC-specific) without examination of effects that are embodied in general software agents. This paper introduces a framework for EC interchanges between general-purpose software agents. Preliminary results are shown that illustrate the EC effects of asynchronous actions of agents within this framework. Building on this framework, coevolutionary agents that interact in a simulated producer/consumer economy are introduced. Using these preliminary results as illustrations, areas for future investigation of embodied EC software agents are discussed.

Learning classifier systms (LCSs) offer a unique opportunity to study the adaptive exploitation of memory. Because memory is manipulated in the form of simple internal messages in the LCS, one can easily and carefully examine the development of a system of internal memory symbols. This study examines the LCS applied to a problem whose only performance goal is the effective exploitation of memory. Experimental results show that the genetic algorithm forms a relatively effective set of internal memory symbols, but that this effectiveness is directly limited by the emergence of parasite rules. The results indicate that the emergence of parasites may be an inevitable consequence in a system that must evolve its own set of internal memory symbols. The paper's primary conclusion is that the emergence of parasites is a fundamental obstacle in such problems. To overcome this obstacle, it is suggested that the LCS must form larger, multirule structures. In such structures, parasites can be more accurately evaluated and thus eliminated. This effect is demonstrated through a preliminary evaluation of a classifier corporation scheme. Final comments present future directions for research on memory exploitation in the LCS and similar evolutionary computing systems.

This paper suggests a simple analogy between learning classifier systems (LCSs) and neural networks (NNs). By clarifying the relationship between LCSs and NNs, the paper indicates how techniques from one can be utilized in the other. The paper points out that the primary distinguishing characteristic of the LCS is its use of a co-adaptive genetic algorithm (GA), where the end product of evolution is a diverse population of individuals that cooperate to perform useful computation. This stands in contrast to typical GA/NN schemes, where a population of networks is employed to evolve a single, optimized network. To fully illustrate the LCS/NN analogy used in this paper, an LCS-like NN is implemented and tested. The test is constructed to run parallel to a similar GA/NN study that did not employ a co-adaptive GA. The test illustrates the LCS/NN analogy and suggests an interesting new method for applying GAs in NNs. Final comments discuss extensions of this work and suggest how LCS and NN studies can further benefit each other.

This paper describes an immune system model based on binary strings. The purpose of the model is to study the pattern-recognition processes and learning that take place at both the individual and species levels in the immune system. The genetic algorithm (GA) is a central component of the model. The paper reports simulation experiments on two pattern-recognition problems that are relevant to natural immune systems. Finally, it reviews the relation between the model and explicit fitness-sharing techniques for genetic algorithms, showing that the immune system model implements a form of implicit fitness sharing.

In typical applications, genetic algorithms (GAs) process populations of potential problem solutions to evolve a single population member that specifies an ‘optimized’ solution. The majority of GA analysis has focused on these optimization applications. In other applications (notably learning classifier systems and certain connectionist learning systems), a GA searches for a population of cooperative structures that jointly perform a computational task. This paper presents an analysis of this type of GA problem. The analysis considers a simplified genetics-based machine learning system: a model of an immune system. In this model, a GA must discover a set of pattern-matching antibodies that effectively match a set of antigen patterns. Analysis shows how a GA can automatically evolve and sustain a diverse, cooperative population. The cooperation emerges as a natural part of the antigen-antibody matching procedure. This emergent effect is shown to be similar to fitness sharing, an explicit technique for multimodal GA optimization. Further analysis shows how the GA population can adapt to express various degrees of generalization. The results show how GAs can automatically and simultaneously discover effective groups of cooperative computational structures.