Sparse military satellite constellations were designed using two methods: a traditional approach and a genetic algorithm. One of the traditional constellation designs was the Discoverer II space based radar. Discoverer II was an 8 plane, 24 satellite, Low Earth Orbit (LEO), Walker constellation designed to provide high-range resolution ground moving target indication (HRR-GMTI), synthetic aperture radar (SAR) imaging and high resolution digital terrain mapping. The traditional method designed 9-ball, 12-ball, 18-ball, and 24- ball Walker constellations. The genetic algorithm created constellations by deriving a phenotype from a triploid genotype encoding of orbital elements. The performance of both design methods were compared using a computer simulation. The fitness of each constellation was calculated using maximum gap time, maximum revisit time, and percent coverage. The goal was to determine if one design method would consistently outperform the other. The genetic algorithm offered a fitness improvement over traditional constellation design methods in all cases except the 24-ball constellation where it demonstrated comparable results. The genetic algorithm improvement over the traditional constellations increased as the number of satellites per constellation decreased. A derived equation related revisit time to the number of ship tracks maintained. / US Navy (USN) author.

This thesis presents an implementation of a learning classifier system which learns good database schema. The system is implemented in Java using the NetBeans development environment, which provides a good control for the GUI components. The system contains four components: a user interface, a rule and message system, an apportionment of credit system, and genetic algorithms. The input of the system is a set of simple database schemas and the objective for the classifier system is to keep the good database schemas which are represented by classifiers. The learning classifier system is given some basic knowledge about database concepts or rules. The result showed that the system could decrease the bad schemas and keep the good ones.

Despite considerable advances in missing data imputation techniques over the last three decades, the problem of missing data remains largely unsolved. Many techniques have emerged in the literature as candidate solutions, including the Expectation Maximisation (EM), and the combination of autoassociative neural networks and genetic algorithms (NN-GA). The merits of both these techniques have been discussed at length in the literature, but have never been compared to each other. This thesis contributes to knowledge by firstly, conducting a comparative study of these two techniques.. The significance of the difference in performance of the methods is presented. Secondly, predictive analysis methods suitable for the missing data problem are presented. The predictive analysis in this problem is aimed at determining if data in question are predictable and hence, to help in choosing the estimation techniques accordingly. Thirdly, a novel treatment of missing data for online condition monitoring problems is presented. An ensemble of three autoencoders together with hybrid Genetic Algorithms (GA) and fast simulated annealing was used to approximate missing data. Several significant insights were deduced from the simulation results. It was deduced that for the problem of missing data using computational intelligence approaches, the choice of optimisation methods plays a significant role in prediction. Although, it was observed that hybrid GA and Fast Simulated Annealing (FSA) can converge to the same search space and to almost the same values they differ significantly in duration. This unique contribution has demonstrated that a particular interest has to be paid to the choice of optimisation techniques and their decision boundaries. iii Another unique contribution of this work was not only to demonstrate that a dynamic programming is applicable in the problem of missing data, but to also show that it is efficient in addressing the problem of missing data. An NN-GA model was built to impute missing data, using the principle of dynamic programing. This approach makes it possible to modularise the problem of missing data, for maximum efficiency. With the advancements in parallel computing, various modules of the problem could be solved by different processors, working together in parallel. Furthermore, a method for imputing missing data in non-stationary time series data that learns incrementally even when there is a concept drift is proposed. This method works by measuring the heteroskedasticity to detect concept drift and explores an online learning technique. New direction for research, where missing data can be estimated for nonstationary applications are opened by the introduction of this novel method. Thus, this thesis has uniquely opened the doors of research to this area. Many other methods need to be developed so that they can be compared to the unique existing approach proposed in this thesis. Another novel technique for dealing with missing data for on-line condition monitoring problem was also presented and studied. The problem of classifying in the presence of missing data was addressed, where no attempts are made to recover the missing values. The problem domain was then extended to regression. The proposed technique performs better than the NN-GA approach, both in accuracy and time efficiency during testing. The advantage of the proposed technique is that it eliminates the need for finding the best estimate of the data, and hence, saves time. Lastly, instead of using complicated techniques to estimate missing values, an imputation approach based on rough sets is explored. Empirical results obtained using both real and synthetic data are given and they provide a valuable and promising insight to the problem of missing data. The work, has significantly confirmed that rough sets can be reliable for missing data estimation in larger and real databases.

The quantitative study of conflict management is concerned with finding models which are accurate and also capable of providing a causal interpretation of results. This dissertation applies computational intelligence methods to study interstate disputes. Both multilayer perceptron neural networks and Takagi-Sugeno neuro-fuzzy models are used to model interstate interactions. The multilayer perceptron neural network is trained in the Bayesian framework, using the Hybrid Monte Carlo method to sample from the posterior probabilities. It is found that the network is able to forecast conflict with an accuracy of 77.3%. A hybrid machine learning method using the neural network and the genetic algorithm is then presented as a method of suggesting how conflict can be brought under control. The automatic relevance determination approach and the sensitivity analysis are used as methods of extracting causal information from the neural network. The Takagi-Sugeno neuro-fuzzy model is optimised, using the Gustafson-Kessel clustering algorithm to partion the input space. It is found that the neuro-fuzzy model predicts conflict with an accuracy of 80.1%. The neuro-fuzzy model is also incorporated into the hybrid machine learning method to suggest how the identified conflict cases can be avoided. The casual interpretation is then formulated by a linguistic approximation of the fuzzy rules extracted from the neuro-fuzzy model. The major finding in this work is that the interpretations drawn from both the neural network and the neuro-fuzzy model are consistent.

Na Engenharia Civil são diversos os métodos aplicados visando à otimização de estruturas. Esta dissertação apresenta um estudo e uma aplicação de um desses métodos: os Algoritmos Genéticos (AG\'s). Os Algoritmos Genéticos são algoritmos de busca, não-determinísticos, que trabalham com amostras do conjunto de soluções e se inspiram na teoria da evolução das espécies para resolver o problema. Neste trabalho de pesquisa buscou-se apresentar as principais técnicas e parâmetros utilizados por diversos autores neste tema. Como objetivo principal pretendeu-se, através dos conhecimentos adquiridos sobre o assunto, aplicá-lo na otimização topológica e paramétrica de vigas de concreto armado, submetidas a um carregamento distribuído. Adotaram-se restrições laterais das variáveis e comportamentais (tensões máximas admissíveis - ELU). Procurou-se trabalhar com variáveis discretas, que melhor representam a realidade do projetista de estruturas. Para aplicação desta técnica implementou-se um programa, em linguagem Java seguindo o paradigma de programação orientada a objetos. O programa foi testado aplicando-se a um problema de otimização abordado por outros autores. Um deles utilizou uma abordagem determinística para a solução do problema. Outro utilizou uma abordagem probabilística, porém com variáveis contínuas. Em 85% dos casos o programa (nomeado AGEN) conseguiu encontrar a solução ótima. Concluiu-se que os algoritmos genéticos são uma técnica bastante robusta, que proporciona resultados significativos, principalmente quando se trata de problemas complexos, com variáveis discretas e restrições em constantes mudanças. As deficiências desta técnica são a sua grande dependência em relação à amostra inicial da população, o seu custo computacional e a calibração de parâmetros. Procurou-se, através deste trabalho, apresentar aos pesquisadores e projetistas do campo da engenharia mais uma ferramenta que utiliza técnicas computacionais para encontrar melhores soluções para otimização de estruturas. Pretendendo-se, assim, estimular o desenvolvimento de mais pesquisas sobre este tema bastante promissor. / This work presents a study and application using Genetic Algorithms (GAs) to solve problems that optimization structures, more specifically concrete beans. The GAs are search algorithms, non-deterministics that works with a population of solutions. Its inspired on the evolutions theory of the species to solve problems. In this dissertation sought to show the most used techniques and parameters about this subject. The primary objective was (through the knowledge obtained during this research) to apply it in the topological and parametrical optimization of concrete beams, submitted by a distributed load. Lateral and behavioral constraineds are used. It was tried to work with a discrete variables, which represent more really the context of structures designer. To apply this technique a program was implemented, using the Java language through the oriented object paradigm. The program was tested applying a optimization problem approached by other authors. One of them used a deterministic approach to solution the problem. Another used a probabilistic approach, but with continuous variable. In 85% of the cases the program (called AGEN) get success. It was concluded that genetic algorithms are a very robust technique, which provides significant results, especially in complex problems with discrete variables and constraints on dynamic changes. The weaknesses of this technique are the high dependence on initial population, its computational cost and the parameters calibration. It was, in this work, presenting to scientists and designers in the structural engineering field another tool that uses computational techniques to find better solutions for structures optimization. It pretended to stimulate the development of more research on this topic enough promising.

O mapeamento genético é constituído por procedimentos experimentais e estatísticos que buscam detectar genes associados à etiologia e regulação de doenças, além de estimar os efeitos genéticos e as localizações genômicas correspondentes. Considerando delineamentos experimentais que envolvem cruzamentos controlados de animais ou plantas, diferentes formulações de modelos de regressão podem ser adotados na identificação de QTLs (do inglês, quantitative trait loci), incluindo seus efeitos principais e possíveis efeitos de interação (epistasia). A dificuldade nestes casos de mapeamento é a comparação de modelos que não necessariamente são encaixados e envolvem um espaço de busca de alta dimensão. Para este trabalho, descrevemos um método geral para melhorar a eficiência computacional em mapeamento simultâneo de múltiplos QTLs e de seus efeitos de interação. A literatura tem usado métodos de busca exaustiva ou busca condicional. Propomos o uso do algoritmo genético para pesquisar o espaço multilocos, sendo este mais útil para genomas maiores e mapas densos de marcadores moleculares. Por meio de estudos de simulações mostramos que a busca baseada no algoritmo genético tem eficiência, em geral, mais alta que aquela de um método de busca condicional e que esta eficiência é comparável àquela de uma busca exaustiva. Na formalização do algoritmo genético pesquisamos o comportamento de parâmetros tais como: probabilidade de recombinação, probabilidade de mutação, tamanho amostral, quantidade de gerações, quantidade de soluções e tamanho do genoma, para diferentes funções objetivo: BIC (do inglês, Bayesian Information Criterion), AIC (do inglês, Akaike Information Criterion) e SSE, a soma de quadrados dos resíduos de um modelo ajustado. A aplicação das metodologias propostas é também considerada na análise de um conjunto de dados genotípicos e fenotípicos de ratos provenientes de um delineamento F2. / Genetic mapping is defined in terms of experimental and statistical procedures applied for detection and localization of genes associated to the etiology and regulation of diseases. Considering experimental designs in controlled crossings of animals or plants, different formulations of regression models can be adopted in the identification of QTL\'s (Quantitative Trait Loci) to the inclusion of the main and interaction effects between genes (epistasis). The difficulty in these approaches of gene mapping is the comparison of models that are not necessarily nested and involves a multiloci search space of high dimension. In this work, we describe a general method to improve the computational efficiency in simultaneous mapping of multiples QTL\'s and their interactions effects. The literature has used methods of exhausting search or conditional search. We consider the genetic algorithm to search the multiloci space, looking for epistatics loci distributed on the genome. Compared to the others procedures, the advantage to use such algorithm increases more for set of genes bigger and dense maps of molecular markers. Simulation studies have shown that the search based on the genetic algorithm has efficiency, in general, higher than the conditional search and that its efficiency is comparable to that one of an exhausting search. For formalization of the genetic algorithm we consider different values of the parameters as recombination probability, mutation probability, sample size, number of generations, number of solutions and size of the set of genes. We evaluate different objective functions under the genetic algorithm: BIC, AIC and SSE. In addition, we used the sample phenotypic and genotypic data bank. Briefly, the study examined blood pressure variation before and after a salt loading experiment in an intercross (F2) progeny.

Este trabalho apresenta a técnica de Algoritmo Genético para apoiar o processo de modelagem e simulação de sistemas complexos. Tal técnica pode ser vista como uma opção as técnicas tradicionais de modelagem e análise dos dados de simulação, simplificando todo esse processo, por suas características de simplicidade e generalidade, não exigindo conhecimento especifico do domínio do problema. É apresentada uma ferramenta computacional - SimAG - baseada em Algoritmo Genético para modelagem e simulação de sistemas dessa natureza. Um exemplo de aplicação e estudado, onde pode ser constatada a viabilidade da utilização da técnica no processo de simulação / This work presents the Genetic Algorithm technique supporting the process of complex systems modeling and simulation. Such technique can be seen as an option to both the traditional modeling and the simulation data analysis techniques. Due to its inherent simplicity and generic application, it simplifies the whole process of simulation, demanding no specific knowledge over the problem domain. A genetic algorithm based computational tool - SimAG - for the modeling and simulation of such systems is here presented, and an example of its use is analyzed, thus demonstrating the feasibility of this new technique application in simulation processes

Os sistemas de distribuição de energia elétrica, sujeitos a um crescimento contínuo de uso, tornam-se cada vez mais complexos e sofisticados na demanda atual de mercado. Neste contexto, surge a necessidade de respostas rápidas para diagnósticos da manutenção preventiva e corretiva da rede de distribuição. Quatro das principais aplicações necessárias são: 1) os cálculos de índices de continuidade, a partir de um circuito elétrico; 2) a definição das topologias com os equipamentos de proteção no circuito; 3) simulação do comportamento do circuito elétrico diante de uma série histórica de contingências de ocorrências na rede; e 4) análise do conjunto dos equipamentos de proteção existentes no inventário da concessionária. Este trabalho tem por objetivo apresentar os estudos preliminares na busca de um modelo de circuito elétrico otimizado, sob os pontos de vista técnicos e operacionais. Em seguida, passa-se à realização e detalhamento dos cálculos apresentados, de acordo com o modelo brasileiro. O uso de uma metodologia para o desenvolvimento das soluções de software, em especial o modelo em cascata, permite desenvolver todos os seus blocos funcionais dentro do processo. Os resultados revelaram que estas soluções, desenvolvidas com inteligência artificial (algoritmos genéticos - AG), alcançam bons resultados quanto ao circuito otimizado, apresentando valores compatíveis quando comparados aos modelos estudados: resultou ainda em um esforço computacional otimizado e boa convergência de valores. A análise do histórico de contingências também retorna o estado final do circuito e os índices de continuidade (DIC, FIC e DMIC). A otimização do uso de um inventário prévio de equipamentos baseada em AG retornou o melhor circuito otimizado, considerando-se a realidade da concessionária. Um aspecto relevante, que pode ser aplicado pelas concessionárias, refere-se aos arquivos de saída: os valores finais dos índices de continuidade e as informações gráficas do circuito otimizado proporcionam uma análise rápida sobre o comportamento do circuito. / The electric power distribution systems, submitted to a continuous growth, become each time more complex and sophisticated in the current demand of market. In this context, appears the need of fast answers for diagnosis of the preventive and corrective maintenance of the distribution network. Four of the main necessary applications are: 1) the calculations of continuity indices, from an electric circuit; 2) the definition of the topologies of the equipments of protection in the circuit; 3) simulation of the behavior of the electric circuit ahead of a historical series of contingencies of occurrences in the network; and 4) analysis of the set of the existing equipment of protection in the inventory of the electric power company. This thesis aims at presenting the preliminary studies in the search of a model of optimized electrical circuit, under the technical and operational view points. After that, it is transferred the accomplishment and detailing of the presented calculations, in accordance with the Brazilian model. The use of a methodology for the development of the solutions of software, in special the waterfall model, allows to develop all its functional blocks inside of the process. The results had disclosed that these solutions, developed with artificial intelligence (genetic algorithms - GA), reach good resulted how much to the optimized circuit, presenting compatible values when compared with the studied models: it still resulted in a computational effort optimized and good convergence of values. The analysis of the description of contingencies also returns the final state from the circuit and the continuity indices (DIC, FIC and DMIC). The optimization of the use of a previous inventory of equipments based on GA returned the best optimized circuit, considering the reality of the electric power company. An excellent aspect, that can be applied by the electric power companies, makes the reference to the output archives: the final values of the continuity indices and the graphical informations of the optimized circuit provide a fast analysis on the behavior of the circuit.

O componente de planejamento faz parte do projeto de desenvolvimento dos veículos autônomos, e é responsável por gerar rotas para o sistema como um todo. Em aplicações em que o veículo deve visitar pontos em intervalos de tempo pré-determinados, o componente de planejamento se enquadra em um problema de roteirização conhecido da literatura, denominado problema de roteirização de veículos com janelas de tempo. Tal problema é uma generalização do problema clássico de roteirização de veículos classificado no grupo de problemas NP-Hard. Esse trabalho apresenta uma proposta de solução para o problema baseada na metaheurística algoritmo genético. Os cromossomos foram representados pela ordem de atendimento dos clientes sem delimitadores de rota. Para quebrar os cromossomos em rotas, foi utilizado um procedimento adaptado baseado em Prins (2004). A população inicial se constitui por uma parte construída com cromossomos criados aleatoriamente e outra parte construída através da heurística de inserção I1 de Solomon (1987), com quatro formas diferentes de inserir o primeiro cliente de cada rota. Na fase de recombinação, foram utilizados quatro tipos de crossover: uniforme, dois pontos, heurístico e PMX, e um operador de mutação baseado em uma busca heurística. A cada geração foram aplicados princípios de elitismo e pós-otimização utilizando a heurística -interchange de Osman (1993). O algoritmo foi testado nos conjuntos C1, C2, R1, R2, RC1 e RC2 de Solomon (1987) e os resultados foram comparados com os melhores resultados encontrados na literatura. / The planning component is a part of autonomous vehicle development project and it is responsible to generate routes for the system as a whole. In applications which vehicle must to visit way points at predetermined intervals of time, the planning component fits into a routing problem known in the literature called routing problem with time windows. This problem is a generalization of the classical vehicle routing problem classified in the group of NP- Hard problems. This thesis presents a solution proposal to problem based on genetic algorithm metaheuristic. Chromosomes were represented by the order of serving customers without delimiters route. To split the chromosomes on routes, it is used a procedure adapted based on Prins (2004). The initial population is constituted by two parts: one with randomly created chromosomes and another constructed through the insertion heuristic I1 of Solomon (1987), with four different ways of insertion of the first customer of each route. In the recombination step, four types of crossover were used: uniform, two points, heuristic, and PMX, and a mutation operator based on heuristic search. In each generation it is applied principles of elitism and postoptimization using the -interchange heuristic of Osman (1993). The algorithm was tested on the sets C1, C2, R1, R2, RC1 and RC2 of Solomon (1987) and the results were compared with the best results found in the literature.

by Wong Man Leung. / Thesis (Ph.D.)--Chinese University of Hong Kong, 1995. / Includes bibliographical references (leaves 227-236). / List of Figures --- p.iii / List of Tables --- p.vi / Chapter Chapter 1 : --- Introduction --- p.1 / Chapter 1.1. --- Automatic programming and program induction --- p.1 / Chapter 1.2. --- Motivation --- p.6 / Chapter 1.3. --- Contributions of the research --- p.8 / Chapter 1.4. --- Outline of the thesis --- p.11 / Chapter Chapter 2 : --- An Overview of Evolutionary Algorithms --- p.13 / Chapter 2.1. --- Evolutionary algorithms --- p.13 / Chapter 2.2. --- Genetic Algorithms (GAs) --- p.15 / Chapter 2.2.1. --- The canonical genetic algorithm --- p.16 / Chapter 2.2.1.1. --- Selection methods --- p.21 / Chapter 2.2.1.2. --- Recombination methods --- p.24 / Chapter 2.2.1.3. --- Inversion and Reordering --- p.27 / Chapter 2.2.2. --- Implicit parallelism and the building block hypothesis --- p.28 / Chapter 2.2.3. --- Steady state genetic algorithms --- p.32 / Chapter 2.2.4. --- Hybrid algorithms --- p.33 / Chapter 2.3. --- Genetic Programming (GP) --- p.34 / Chapter 2.3.1. --- Introduction to the traditional GP --- p.34 / Chapter 2.3.2. --- Automatic Defined Function (ADF) --- p.41 / Chapter 2.3.3. --- Module Acquisition (MA) --- p.44 / Chapter 2.3.4. --- Strongly Typed Genetic Programming (STGP) --- p.49 / Chapter 2.4. --- Evolution Strategies (ES) --- p.50 / Chapter 2.5. --- Evolutionary Programming (EP) --- p.55 / Chapter Chapter 3 : --- Inductive Logic Programming --- p.59 / Chapter 3.1. --- Inductive concept learning --- p.59 / Chapter 3.2. --- Inductive Logic Programming (ILP) --- p.62 / Chapter 3.2.1. --- Interactive ILP --- p.64 / Chapter 3.2.2. --- Empirical ILP --- p.65 / Chapter 3.3. --- Techniques and methods of ILP --- p.67 / Chapter Chapter 4 : --- Genetic Logic Programming and Applications --- p.74 / Chapter 4.1. --- Introduction --- p.74 / Chapter 4.2. --- Representations of logic programs --- p.76 / Chapter 4.3. --- Crossover of logic programs --- p.81 / Chapter 4.4. --- Genetic Logic Programming System (GLPS) --- p.87 / Chapter 4.5. --- Applications --- p.90 / Chapter 4.5.1. --- The Winston's arch problem --- p.91 / Chapter 4.5.2. --- The modified Quinlan's network reachability problem --- p.92 / Chapter 4.5.3. --- The factorial problem --- p.95 / Chapter Chapter 5 : --- The logic grammars based genetic programming system (LOGENPRO) --- p.100 / Chapter 5.1. --- Logic grammars --- p.101 / Chapter 5.2. --- Representations of programs --- p.103 / Chapter 5.3. --- Crossover of programs --- p.111 / Chapter 5.4. --- Mutation of programs --- p.126 / Chapter 5.5. --- The evolution process of LOGENPRO --- p.130 / Chapter 5.6. --- Discussion --- p.132 / Chapter Chapter 6 : --- Applications of LOGENPRO --- p.134 / Chapter 6.1. --- Learning functional programs --- p.134 / Chapter 6.1.1. --- Learning S-expressions using LOGENPRO --- p.134 / Chapter 6.1.2. --- The DOT PRODUCT problem --- p.137 / Chapter 6.1.2. --- Learning sub-functions using explicit knowledge --- p.143 / Chapter 6.2. --- Learning logic programs --- p.148 / Chapter 6.2.1. --- Learning logic programs using LOGENPRO --- p.148 / Chapter 6.2.2. --- The Winston's arch problem --- p.151 / Chapter 6.2.3. --- The modified Quinlan's network reachability problem --- p.153 / Chapter 6.2.4. --- The factorial problem --- p.154 / Chapter 6.2.5. --- Discussion --- p.155 / Chapter 6.3. --- Learning programs in C --- p.155 / Chapter Chapter 7 : --- Knowledge Discovery in Databases --- p.159 / Chapter 7.1. --- Inducing decision trees using LOGENPRO --- p.160 / Chapter 7.1.1. --- Decision trees --- p.160 / Chapter 7.1.2. --- Representing decision trees as S-expressions --- p.164 / Chapter 7.1.3. --- The credit screening problem --- p.166 / Chapter 7.1.4. --- The experiment --- p.168 / Chapter 7.2. --- Learning logic program from imperfect data --- p.174 / Chapter 7.2.1. --- The chess endgame problem --- p.177 / Chapter 7.2.2. --- The setup of experiments --- p.178 / Chapter 7.2.3. --- Comparison of LOGENPRO with FOIL --- p.180 / Chapter 7.2.4. --- Comparison of LOGENPRO with BEAM-FOIL --- p.182 / Chapter 7.2.5. --- Comparison of LOGENPRO with mFOILl --- p.183 / Chapter 7.2.6. --- Comparison of LOGENPRO with mFOIL2 --- p.184 / Chapter 7.2.7. --- Comparison of LOGENPRO with mFOIL3 --- p.185 / Chapter 7.2.8. --- Comparison of LOGENPRO with mFOIL4 --- p.186 / Chapter 7.2.9. --- Comparison of LOGENPRO with mFOIL5 --- p.187 / Chapter 7.2.10. --- Discussion --- p.188 / Chapter 7.3. --- Learning programs in Fuzzy Prolog --- p.189 / Chapter Chapter 8 : --- An Adaptive Inductive Logic Programming System --- p.192 / Chapter 8.1. --- Adaptive Inductive Logic Programming --- p.192 / Chapter 8.2. --- A generic top-down ILP algorithm --- p.196 / Chapter 8.3. --- Inducing procedural search biases --- p.200 / Chapter 8.3.1. --- The evolution process --- p.201 / Chapter 8.3.2. --- The experimentation setup --- p.202 / Chapter 8.3.3. --- Fitness calculation --- p.203 / Chapter 8.4. --- Experimentation and evaluations --- p.204 / Chapter 8.4.1. --- The member predicate --- p.205 / Chapter 8.4.2. --- The member predicate in a noisy environment --- p.205 / Chapter 8.4.3. --- The multiply predicate --- p.206 / Chapter 8.4.4. --- The uncle predicate --- p.207 / Chapter 8.5. --- Discussion --- p.208 / Chapter Chapter 9 : --- Conclusion and Future Work --- p.210 / Chapter 9.1. --- Conclusion --- p.210 / Chapter 9.2. --- Future work --- p.217 / Chapter 9.2.1. --- Applying LOGENPRO to discover knowledge from databases --- p.217 / Chapter 9.2.2. --- Learning recursive programs --- p.218 / Chapter 9.2.3. --- Applying LOGENPRO in engineering design --- p.220 / Chapter 9.2.4. --- Exploiting parallelism of evolutionary algorithms --- p.222 / Reference --- p.227 / Appendix A --- p.237