No description available.

No description available.

Artificial neural networks learned by evolutionary algorithms are commonly used to control the robots. Neural networks can be encoded either directly as a list of weights or indirectly as a weight generator. Unlike direct coding indirect encoding allows to encode a large network using a short genetic code. HyperNEAT is a neuroevolutionary algorithm, which encodes the neural network indirectly, through another (producing) network, which computes synaptic weights. A different algorithm called HyperGP is an alternative to HyperNEAT. In HyperGP, the producing network is replaced by an arithmetic expression, which is being evolved using a genetic programming (GP). We have designed enhancements for HyperGP, using techniques that are either known in a different context of GP or completely new. Algorithm and enhancements have been implemented and experimentally tested on a task of controlling virtual walking robot. The results were compared with HyperNEAT and with the original HyperGP. We have shown that most of the proposed enhancements are effective and, on the given task, HyperGP is better than HyperNEAT. GP thus can successfully replace NEAT in hyper-encoding scheme and improve its efficiency. Powered by TCPDF (www.tcpdf.org)

This research applies the biologically inspired, artificial evolutionary processes of Genetic Algorithms and Genetic Programming to digital hardware circuit synthesis and minimization. In this new application, three approaches are taken to genetic hardware development. First, as a method for logic synthesis, Genetic Programming is applied to the building of logic functions. Experimental results have shown the logic equations from this technique produce better than 88% coverage of the given truth-tables, but the method cannot guarantee complete (100%) coverage. Secondly, to better achieve complete function coverage, an XOR Correction Circuit Algorithm used in conjunction with the Genetic Logic Synthesis was developed. With this algorithm, the genetic logic synthesis can reiteratively attempt coverage by formulating its own selective "correction" functions, for input combinations where complete truth table coverage has not previously been achieved. With this technique, complete function coverage was synthesized in all experiments conducted. The third application of the paradigm is to the minimization of Reed-Muller Equations. In this application, a Genetic Algorithm is implemented only in the search space of all "correct", functionally equivalent equations, with only the task of finding reductions. With this limited search space the solutions have absolute guaranteed function coverage, as well as a better defined focus for the genetic evolutionary process. In both the logic synthesis and minimization processes the genetic operators determine efficient circuit implementations and reductions. The results are often different from those of human designers. Because the genetic techniques incorporate logical testing into the design and build process, one can be assured that the circuit will function as derived on completion. For all three applications, the effects of a number of evolutionary parameters on the genetic operators' problem solving capability are examined. The resulting logic and logic minimizations are also compared with both arbitrarily defined functions and well known logic synthesis benchmarks. It has been shown that genetic operators applied to digital logic can effectively find good solutions for both logic synthesis and logic minimization. / Graduation date: 1997

To locate and to retrieve the needed information from the Internet is an important issue. Existing search engines may give too much useless and redundancy information. Due to the search feature is different for different search engines, it¡¦s very difficult to find an optimal search scheme for all subjects. In this paper, we propose a genetic programming web search system (GPWS) to generate exact query according to a user¡¦s interests. The system can retrieve the information from the search engines, filter the retrieved results and remove the redundancy and useless results. The filtered results are displayed on a uniform user interface. Compared with the queries generated by randomly, the degree of similarity of results and user¡¦s interests are improved.

Thesis (M. Phil.)--University of Hong Kong, 2001. / Includes bibliographical references (leaves 71-76).

Classification under streaming data constraints implies that training must be performed continuously, can only access individual exemplars for a short time after they arrive, must adapt to dynamic behaviour over time, and must be able to retrieve a current classifier at any time. A coevolutionary genetic programming framework is adapted to operate in non-stationary streaming data environments. Methods to generate synthetic datasets for benchmarking streaming classification algorithms are introduced, and the proposed framework is evaluated against them. The use of Pareto archiving is evaluated as a mechanism for retaining access to a limited number of useful exemplars throughout training, and several fitness sharing heuristics for archiving are evaluated. Fitness sharing alone is found to be most effective under streams with continuous (incremental) changes, while the addition of an aging heuristic is preferred when the stream has stepwise changes. Tapped delay lines are explored as a method for explicitly incorporating sequence context in cyclical data streams, and their use in combination with the aging heuristic suggests a promising route forward. / Hyperref'd copy available at: https://web.cs.dal.ca/~atwater/

Classification under streaming data conditions requires that the machine learning approach operate interactively with the stream content. Thus, given some initial machine learning classification capability, it is not possible to assume that the process `generating' stream content will be stationary. It is therefore necessary to first detect when the stream content changes. Only after detecting a change, can classifier retraining be triggered. Current methods for change detection tend to assume an entropy filter approach, where class labels are necessary. In practice, labeling the stream would be extremely expensive. This work proposes an approach in which the behavior of GP individuals is used to detect change without} the use of labels. Only after detecting a change is label information requested. Benchmarking under three computer network traffic analysis scenarios demonstrates that the proposed approach performs at least as well as the filter method, while retaining the advantage of requiring no labels.

This thesis proposes two new variations on the genetic algorithm. The first attempts to improve clustering problems by optimizing the structure of a genetic string dynamically during the run of the algorithm. This is done by using a permutation on the allele which is inherited by the next generation. The second is a multiple pool technique which ensures continuing convergence by maintaining unique lineages and merging pools of similar age. These variations will be tested against two well-known graph theory problems, the Traveling Salesman Problem and the Maximum Clique Problem. The results will be analyzed with respect to string rates, child improvement, pool rating resolution, and average string age. / Department of Computer Science

Thesis (M.S.)--Rochester Institute of Technology, 2009. / Typescript. Includes bibliographical references (leaves 46-47).