• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 20
  • 19
  • 6
  • 2
  • 2
  • 1
  • 1
  • Tagged with
  • 56
  • 29
  • 21
  • 18
  • 17
  • 16
  • 14
  • 12
  • 11
  • 11
  • 10
  • 9
  • 9
  • 9
  • 8
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
11

Adapting Evolutionary Approaches for Optimization in Dynamic Environments

Younes, Abdunnaser January 2006 (has links)
Many important applications in the real world that can be modelled as combinatorial optimization problems are actually dynamic in nature. However, research on dynamic optimization focuses on continuous optimization problems, and rarely targets combinatorial problems. Moreover, dynamic combinatorial problems, when addressed, are typically tackled within an application context. <br /><br /> In this thesis, dynamic combinatorial problems are addressed collectively by adopting an evolutionary based algorithmic approach. On the plus side, their ability to manipulate several solutions at a time, their robustness and their potential for adaptability make evolutionary algorithms a good choice for solving dynamic problems. However, their tendency to converge prematurely, the difficulty in fine-tuning their search and their lack of diversity in tracking optima that shift in dynamic environments are drawbacks in this regard. <br /><br /> Developing general methodologies to tackle these conflicting issues constitutes the main theme of this thesis. First, definitions and measures of algorithm performance are reviewed. Second, methods of benchmark generation are developed under a generalized framework. Finally, methods to improve the ability of evolutionary algorithms to efficiently track optima shifting due to environmental changes are investigated. These methods include adapting genetic parameters to population diversity and environmental changes, the use of multi-populations as an additional means to control diversity, and the incorporation of local search heuristics to fine-tune the search process efficiently. <br /><br /> The methodologies developed for algorithm enhancement and benchmark generation are used to build and test evolutionary models for dynamic versions of the travelling salesman problem and the flexible manufacturing system. Results of experimentation demonstrate that the methods are effective on both problems and hence have a great potential for other dynamic combinatorial problems as well.
12

Concurrent design of facility layout and flow-based department formation

Chae, Junjae 17 February 2005 (has links)
The design of facility layout takes into account a number of issues including the formation of departments, the layout of these, the determination of the material handling methods to be used, etc. To achieve an e&#64259;cient layout, these issues should be examined simultaneously. However, in practice, these problems are generally formulated and solved sequentially due to the complicated nature of the integrated problem. Speci&#64257;cally, there is close interaction between the formation of departments and layout of these departments. These problems are treated as separate problems that are solved sequentially. This procedure is mainly due to the complexity of each problem and the interrelationships between them. In this research, we take a &#64257;rst step toward integrating the &#64258;ow-based department formation and departmental layout into comprehensive mathematical models and develop appropriate solution procedures. It is expected that these mathematical models and the solution procedures developed will generate more e&#64259;cient manufacturing system designs, insights into the nature of the concurrent facility layout problem, and new research directions.
13

Memetic Algorithms For Timetabling Problems In Private Schools

Aldogan, Deniz 01 July 2005 (has links) (PDF)
The aim of this study is to introduce a real-world timetabling problem that exists in some private schools in Turkey and to solve such problem instances utilizing memetic algorithms. Being a new type of problem and for privacy reasons, there is no real data available. Hence for benchmarking purposes, a random data generator has been implemented. Memetic algorithms (MAs) combining genetic algorithms and hill-climbing are applied to solve synthetic problem instances produced by this generator. Different types of recombination and mutation operators based on the hierarchical structure of the timetabling problem are proposed. A modified version of the violation directed hierarchical hill-climbing method (VDHC), introduced by A. Alkan and E. Ozcan, coordinates the process of 12 different low-level hill-climbing operators grouped in two distinct arrangements that attempt to resolve corresponding constraint violations. VDHC is an adaptive method advocating cooperation of hill-climbing operators. In addition, MAs with VDHC are compared with different versions of multimeme algorithms and pure genetic algorithms. Experimental results on synthetic benchmark data set indicate the success of the proposed MA.
14

Uma proposta de algoritmo memético baseado em conhecimento para o problema de predição de estruturas 3-D de proteínas

Correa, Leonardo de Lima January 2017 (has links)
Algoritmos meméticos são meta-heurísticas evolutivas voltadas intrinsecamente à exploração e incorporação de conhecimentos relacionados ao problema em estudo. Nesta dissertação, foi proposto um algoritmo memético multi populacional baseado em conhecimento para lidar com o problema de predição de estruturas tridimensionais de proteínas voltado à modelagem de estruturas livres de similaridades conformacionais com estruturas de proteínas determinadas experimentalmente. O algoritmo em questão, foi estruturado em duas etapas principais de processamento: (i) amostragem e inicialização de soluções; e (ii) otimização dos modelos estruturais provenientes da etapa anterior. A etapa I objetiva a geração e classificação de diversas soluções, a partir da estratégia Lista de Probabilidades Angulares, buscando a definição de diferentes grupos estruturais e a criação de melhores estruturas a serem incorporadas à meta-heurística como soluções iniciais das multi populações. A segunda etapa consiste no processo de otimização das estruturas oriundas da etapa I, realizado por meio da aplicação do algoritmo memético de otimização, o qual é fundamentado na organização da população de indivíduos em uma estrutura em árvore, onde cada nodo pode ser interpretado como uma subpopulação independente, que ao longo do processo interage com outros nodos por meio de operações de busca global voltadas a características do problema, visando o compartilhamento de informações, a diversificação da população de indivíduos, e a exploração mais eficaz do espaço de busca multimodal do problema O algoritmo engloba ainda uma implementação do algoritmo colônia artificial de abelhas, com o propósito de ser utilizado como uma técnica de busca local a ser aplicada em cada nodo da árvore. O algoritmo proposto foi testado em um conjunto de 24 sequências de aminoácidos, assim como comparado a dois métodos de referência na área de predição de estruturas tridimensionais de proteínas, Rosetta e QUARK. Os resultados obtidos mostraram a capacidade do método em predizer estruturas tridimensionais de proteínas com conformações similares a estruturas determinadas experimentalmente, em termos das métricas de avaliação estrutural Root-Mean-Square Deviation e Global Distance Total Score Test. Verificou-se que o algoritmo desenvolvido também foi capaz de atingir resultados comparáveis ao Rosetta e ao QUARK, sendo que em alguns casos, os superou. Corroborando assim, a eficácia do método. / Memetic algorithms are evolutionary metaheuristics intrinsically concerned with the exploiting and incorporation of all available knowledge about the problem under study. In this dissertation, we present a knowledge-based memetic algorithm to tackle the threedimensional protein structure prediction problem without the explicit use of template experimentally determined structures. The algorithm was divided into two main steps of processing: (i) sampling and initialization of the algorithm solutions; and (ii) optimization of the structural models from the previous stage. The first step aims to generate and classify several structural models for a determined target protein, by the use of the strategy Angle Probability List, aiming the definition of different structural groups and the creation of better structures to initialize the initial individuals of the memetic algorithm. The Angle Probability List takes advantage of structural knowledge stored in the Protein Data Bank in order to reduce the complexity of the conformational search space. The second step of the method consists in the optimization process of the structures generated in the first stage, through the applying of the proposed memetic algorithm, which uses a tree-structured population, where each node can be seen as an independent subpopulation that interacts with others, over global search operations, aiming at information sharing, population diversity, and better exploration of the multimodal search space of the problem The method also encompasses ad-hoc global search operators, whose objective is to increase the exploration capacity of the method turning to the characteristics of the protein structure prediction problem, combined with the Artificial Bee Colony algorithm to be used as a local search technique applied to each node of the tree. The proposed algorithm was tested on a set of 24 amino acid sequences, as well as compared with two reference methods in the protein structure prediction area, Rosetta and QUARK. The results show the ability of the method to predict three-dimensional protein structures with similar foldings to the experimentally determined protein structures, regarding the structural metrics Root-Mean-Square Deviation and Global Distance Total Score Test. We also show that our method was able to reach comparable results to Rosetta and QUARK, and in some cases, it outperformed them, corroborating the effectiveness of our proposal.
15

Memetic networks : problem-solving with social network models / Redes Meméticas: solução de problemas utilizando modelos de redes sociais

Araújo, Ricardo Matsumura de January 2010 (has links)
Sistemas sociais têm se tornado cada vez mais relevantes para a Ciência da Computação em geral e para a Inteligência Artificial em particular. Tal interesse iniciou-se pela necessidade de analisar-se sistemas baseados em agentes onde a interação social destes agentes pode ter um impacto no resultado esperado. Uma tendência mais recente vem da área de Processamento Social de Informações, Computação Social e outros métodos crowdsourced, que são caracterizados por sistemas de computação compostos de pessoas reais, com um forte componente social na interação entre estas. O conjunto de todas interações sociais e os atores envolvidos compõem uma rede social, que pode ter uma forte influência em o quão eficaz ou eficiente o sistema pode ser. Nesta tese, exploramos o papel de estruturas de redes em sistemas sociais que visam a solução de problemas. Enquadramos a solução de problemas como uma busca por soluções válidas em um espaço de estados e propomos um modelo - a Rede Memética - que é capaz de realizar busca utilizando troca de informações (memes) entre atores interagindo em uma rede social. Tal modelo é aplicado a uma variedade de cenários e mostramos como a presença da rede social pode melhorar a capacidade do sistema em encontrar soluções. Adicionalmente, relacionamos propriedades específicas de diversas redes bem conhecidas ao comportamento observado para os algoritmos propostos, resultando em um conjunto de regras gerais que podem melhorar o desempenho de tais sistemas sociais. Por fim, mostramos que os algoritmos propostos são competitivos com técnicas tradicionais de busca heurística em diversos cenários. / Social systems are increasingly relevant to computer science in general and artificial intelligence in particular. Such interest was first sparkled by agent-based systems where the social interaction of such agents can be relevant to the outcome produced. A more recent trend comes from the general area of Social Information Processing, Social Computing and other crowdsourced systems, which are characterized by computing systems composed of people and strong social interactions between them. The set of all social interactions and actors compose a social network, which may have strong influence on how effective the system can be. In this thesis, we explore the role of network structure in social systems aiming at solving problems, focusing on numerical and combinatorial optimization. We frame problem solving as a search for valid solutions in a state space and propose a model - the Memetic Network - that is able to perform search by using the exchange of information, named memes, between actors interacting in a social network. Such model is applied to a variety of scenarios and we show that the presence of a social network greatly improves the system capacity to find good solutions. In addition, we relate specific properties of many well-known networks to the behavior displayed by the proposed algorithms, resulting in a set of general rules that may improve the performance of such social systems. Finally, we show that the proposed algorithms can be competitive with traditional heuristic search algorithms in a number of scenarios.
16

Memetic networks : problem-solving with social network models / Redes Meméticas: solução de problemas utilizando modelos de redes sociais

Araújo, Ricardo Matsumura de January 2010 (has links)
Sistemas sociais têm se tornado cada vez mais relevantes para a Ciência da Computação em geral e para a Inteligência Artificial em particular. Tal interesse iniciou-se pela necessidade de analisar-se sistemas baseados em agentes onde a interação social destes agentes pode ter um impacto no resultado esperado. Uma tendência mais recente vem da área de Processamento Social de Informações, Computação Social e outros métodos crowdsourced, que são caracterizados por sistemas de computação compostos de pessoas reais, com um forte componente social na interação entre estas. O conjunto de todas interações sociais e os atores envolvidos compõem uma rede social, que pode ter uma forte influência em o quão eficaz ou eficiente o sistema pode ser. Nesta tese, exploramos o papel de estruturas de redes em sistemas sociais que visam a solução de problemas. Enquadramos a solução de problemas como uma busca por soluções válidas em um espaço de estados e propomos um modelo - a Rede Memética - que é capaz de realizar busca utilizando troca de informações (memes) entre atores interagindo em uma rede social. Tal modelo é aplicado a uma variedade de cenários e mostramos como a presença da rede social pode melhorar a capacidade do sistema em encontrar soluções. Adicionalmente, relacionamos propriedades específicas de diversas redes bem conhecidas ao comportamento observado para os algoritmos propostos, resultando em um conjunto de regras gerais que podem melhorar o desempenho de tais sistemas sociais. Por fim, mostramos que os algoritmos propostos são competitivos com técnicas tradicionais de busca heurística em diversos cenários. / Social systems are increasingly relevant to computer science in general and artificial intelligence in particular. Such interest was first sparkled by agent-based systems where the social interaction of such agents can be relevant to the outcome produced. A more recent trend comes from the general area of Social Information Processing, Social Computing and other crowdsourced systems, which are characterized by computing systems composed of people and strong social interactions between them. The set of all social interactions and actors compose a social network, which may have strong influence on how effective the system can be. In this thesis, we explore the role of network structure in social systems aiming at solving problems, focusing on numerical and combinatorial optimization. We frame problem solving as a search for valid solutions in a state space and propose a model - the Memetic Network - that is able to perform search by using the exchange of information, named memes, between actors interacting in a social network. Such model is applied to a variety of scenarios and we show that the presence of a social network greatly improves the system capacity to find good solutions. In addition, we relate specific properties of many well-known networks to the behavior displayed by the proposed algorithms, resulting in a set of general rules that may improve the performance of such social systems. Finally, we show that the proposed algorithms can be competitive with traditional heuristic search algorithms in a number of scenarios.
17

Optimisation déterministe et stochastique pour des problèmes de traitement d'images en grande dimension / Deterministic and stochastic optimization for solving large size inverse problems in image processing

Vu, Thi Thanh Xuan 13 November 2017 (has links)
Dans cette thèse on s’intéresse au problème des décompositions canoniques polyadiques de tenseurs d’ordre $N$ potentiellement grands et sous différentes contraintes (non-négativité, aspect creux lié à une possible surestimation du rang du tenseur). Pour traiter ce problème, nous proposons trois nouvelles approches itératives différentes: deux approches déterministes dont une approche proximale, et une approche stochastique. La première approche étend les travaux de thèse de J-P. Royer au cas de tenseurs de dimension $N$. Dans l’approche stochastique, nous considérons pour la première fois dans le domaine des décompositions tensorielles, des algorithmes génétiques (mimétiques) dont principe général repose sur l’évolution d’une population de candidats. Dans le dernier type d’approche, nous avons considéré un algorithme proximal pré-conditionné (le Block-Coordinate Variable Metric Forward-Backward), algorithme fonctionnant par blocs de données avec une matrice de pré-conditionnement liée à chaque bloc et fondé sur deux étapes successives principales : une étape de gradient et une étape proximale. Finalement, les différentes méthodes suggérées sont comparées entre elles et avec d’autres algorithmes classiques de la littérature sur des données synthétiques (à la fois aléatoires ou proches des données observées en spectroscopie de fluorescence) et sur des données expérimentales réelles correspondant à une campagne de surveillance des eaux d’une rivière et visant à la détection d’apparition de polluants. / In this PhD thesis, we consider the problem of the Canonical Polyadic Decomposition (CPD) of potentially large $N$-th order tensors under different constraints (non-negativity, sparsity due to a possible overestimation of the tensor rank, etc.). To tackle such a problem, we propose three new iterative methods: a standard gradient based deterministic approach, a stochastic approach (memetic) and finally a proximal approach (Block-Coordinate Variable Metric Forward-Backward). The first approach extends J-P. Royer's works to the case of non-negative N-th order tensors. In the stochastic approach, genetic (memetic) methods are considered for the first time to solve the CPD problem. Their general principle is based on the evolution of a family of candidates. In the third type of approaches, a proximal algorithm namely the Block-Coordinate Variable Metric Forward-Backward is presented. The algorithm relies on two main steps: a gradient step and a proximal step. The blocks of coordinates naturally correspond to latent matrices. We propose a majorant function as well as a preconditioner with regard to each block. All methods are compared with other popular algorithms of the literature on synthetic (fluorescence spectroscopy like or random) data and on real experimental data corresponding to a water monitoring campaign aiming at detecting the appearance of pollutants.
18

Uma proposta de algoritmo memético baseado em conhecimento para o problema de predição de estruturas 3-D de proteínas

Correa, Leonardo de Lima January 2017 (has links)
Algoritmos meméticos são meta-heurísticas evolutivas voltadas intrinsecamente à exploração e incorporação de conhecimentos relacionados ao problema em estudo. Nesta dissertação, foi proposto um algoritmo memético multi populacional baseado em conhecimento para lidar com o problema de predição de estruturas tridimensionais de proteínas voltado à modelagem de estruturas livres de similaridades conformacionais com estruturas de proteínas determinadas experimentalmente. O algoritmo em questão, foi estruturado em duas etapas principais de processamento: (i) amostragem e inicialização de soluções; e (ii) otimização dos modelos estruturais provenientes da etapa anterior. A etapa I objetiva a geração e classificação de diversas soluções, a partir da estratégia Lista de Probabilidades Angulares, buscando a definição de diferentes grupos estruturais e a criação de melhores estruturas a serem incorporadas à meta-heurística como soluções iniciais das multi populações. A segunda etapa consiste no processo de otimização das estruturas oriundas da etapa I, realizado por meio da aplicação do algoritmo memético de otimização, o qual é fundamentado na organização da população de indivíduos em uma estrutura em árvore, onde cada nodo pode ser interpretado como uma subpopulação independente, que ao longo do processo interage com outros nodos por meio de operações de busca global voltadas a características do problema, visando o compartilhamento de informações, a diversificação da população de indivíduos, e a exploração mais eficaz do espaço de busca multimodal do problema O algoritmo engloba ainda uma implementação do algoritmo colônia artificial de abelhas, com o propósito de ser utilizado como uma técnica de busca local a ser aplicada em cada nodo da árvore. O algoritmo proposto foi testado em um conjunto de 24 sequências de aminoácidos, assim como comparado a dois métodos de referência na área de predição de estruturas tridimensionais de proteínas, Rosetta e QUARK. Os resultados obtidos mostraram a capacidade do método em predizer estruturas tridimensionais de proteínas com conformações similares a estruturas determinadas experimentalmente, em termos das métricas de avaliação estrutural Root-Mean-Square Deviation e Global Distance Total Score Test. Verificou-se que o algoritmo desenvolvido também foi capaz de atingir resultados comparáveis ao Rosetta e ao QUARK, sendo que em alguns casos, os superou. Corroborando assim, a eficácia do método. / Memetic algorithms are evolutionary metaheuristics intrinsically concerned with the exploiting and incorporation of all available knowledge about the problem under study. In this dissertation, we present a knowledge-based memetic algorithm to tackle the threedimensional protein structure prediction problem without the explicit use of template experimentally determined structures. The algorithm was divided into two main steps of processing: (i) sampling and initialization of the algorithm solutions; and (ii) optimization of the structural models from the previous stage. The first step aims to generate and classify several structural models for a determined target protein, by the use of the strategy Angle Probability List, aiming the definition of different structural groups and the creation of better structures to initialize the initial individuals of the memetic algorithm. The Angle Probability List takes advantage of structural knowledge stored in the Protein Data Bank in order to reduce the complexity of the conformational search space. The second step of the method consists in the optimization process of the structures generated in the first stage, through the applying of the proposed memetic algorithm, which uses a tree-structured population, where each node can be seen as an independent subpopulation that interacts with others, over global search operations, aiming at information sharing, population diversity, and better exploration of the multimodal search space of the problem The method also encompasses ad-hoc global search operators, whose objective is to increase the exploration capacity of the method turning to the characteristics of the protein structure prediction problem, combined with the Artificial Bee Colony algorithm to be used as a local search technique applied to each node of the tree. The proposed algorithm was tested on a set of 24 amino acid sequences, as well as compared with two reference methods in the protein structure prediction area, Rosetta and QUARK. The results show the ability of the method to predict three-dimensional protein structures with similar foldings to the experimentally determined protein structures, regarding the structural metrics Root-Mean-Square Deviation and Global Distance Total Score Test. We also show that our method was able to reach comparable results to Rosetta and QUARK, and in some cases, it outperformed them, corroborating the effectiveness of our proposal.
19

Otimização de alocação de chaves em redes de distribuição de energia elétrica / Optimization of switch allocation in power distribution networks

Assis, Laura Silva de, 1983- 25 August 2018 (has links)
Orientadores: Christiano Lyra Filho, Celso Cavellucci / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-25T04:13:51Z (GMT). No. of bitstreams: 1 Assis_LauraSilvade_D.pdf: 3122445 bytes, checksum: 01644f90a086983b8729f81804874faa (MD5) Previous issue date: 2014 / Resumo: Grande parte das falhas em sistemas elétricos de potência ocorrem por consequência de falhas permanentes nas redes de distribuição. Agências reguladoras definem índices de confiabilidade para quantificar e avaliar a qualidade da distribuição de energia. A violação dos limites estabelecidos podem resultar em multas significativas para a distribuidora de energia. Um dos objetivos ao se realizar a instalação de chaves em redes de distribuição é criar a possibilidade de re-energizar a maior quantidade de clientes no menor tempo possível através da transferência de carga para sistemas que não tiveram seu fornecimento de energia interrompido. Esta tese estuda o problema de alocação de chaves (PAC) em sistemas radiais de distribuição de energia elétrica e propõe a instalação otimizada desses dispositivos em locais apropriados das redes, a fim de melhorar a confiabilidade do sistema pela redução do período que os consumidores ficam sem energia. Uma metodologia baseada nos conceitos de algoritmo memético juntamente com uma população estruturada é proposta neste trabalho para alocar chaves seccionadoras e de manobra, manuais e automáticas, com diferentes capacidades. A função objetivo utilizada busca minimizar o custo de alocação das chaves e o custo da energia não distribuída sob restrições de confiabilidade e fluxo de carga em todos os componentes da rede. É apresentado também um estudo multiobjetivo para o PAC, que procura alocar chaves minimizando simultaneamente os custos de instalação das chaves e da energia não distribuída e maximizando a confiabilidade da rede, sob restrições de fluxos. A abordagem proposta para resolver o PAC mono-objetivo também foi utilizada no PAC multiobjetivo, juntamente com o método do ?-restrito. A metodologia proposta tem o seu bom desempenho confirmado por diferentes estudos de casos com redes reais de grande porte localizadas no estado de São Paulo / Abstract: Most failures in electric power systems occur as a result of permanent faults in distribution networks. Regulatory agencies establish reliability standards indices for quantify and evaluate the quality of power distribution. The infringe of established limits can result in costly fines for the utility suppliers. One of the aim when perform the switches allocation in distribution networks is the possibility of re-energize the largest amount of customers in the shortest possible time by transferring load to other power systems which don¿t had their energy supply interrupted. This thesis studies the switch allocation problem (SAP) in radial systems of electrical power distribution and proposes an optimized installation of these devices in appropriate locations of network, in order to improve the reliability system by the reducing of the period that consumers remains without power. A methodology based on the concepts of memetic algorithm with a structured population is proposed in this thesis to allocate sectionalizing and tie switches of different capacities, with manual or automatic operation schemes. The objective function used seeks to minimize the switches allocation and the energy not supplied costs under constraints of reliability and load flow. A Multi-objective study for SAP is presented, to perform the switches allocation seeks minimize simultaneously the switches installation and energy not supplied costs and maximize the network reliability, under constraints of load flow. The proposed approach to solve the SAP monocriteria was also used in SAP multi-criteria along with the ?-constraint method. The proposed methodology has its good performance confirmed by several case studies with real large networks located in the state of São Paulo / Doutorado / Automação / Doutora em Engenharia Elétrica
20

Programmation par contraintes et découverte de motifs sur données séquentielles / Constraint programming for sequential pattern mining

Vigneron, Vincent 08 December 2017 (has links)
Des travaux récents ont montré l’intérêt de la programmation par contraintes pour la fouille de données. Dans cette thèse, nous nous intéressons à la recherche de motifs sur séquences, et en particulier à la caractérisation, à l’aide de motifs, de classes de séquences pré-établies. Nous proposons à cet effet un langage de modélisation à base de contraintes qui suppose une représentation matricielle du jeu de séquences. Un motif s’y définit comme un ensemble de caractères (ou de patrons) et pour chacun une localisation dans différentes séquences. Diverses contraintes peuvent alors s’appliquer : validité des localisations, couverture d’une classe de séquences, ordre sur les localisations des caractères commun aux séquences, etc. Nous formulons deux problèmes de caractérisation NP-complets : la caractérisation par motif totalement ordonné (e.g. sous-séquence exclusive à une classe) ou partiellement ordonné. Nous en donnons deux modélisations CSP qui intègrent des contraintes globales pour la preuve d’exclusivité. Nous introduisons ensuite un algorithme mémétique pour l’extraction de motifs partiellement ordonnés qui s’appuie sur la résolution CSP lors des phases d’initialisation et d’intensification. Cette approche hybride se révèle plus performante que l’approche CSP pure sur des séquences biologiques. La mise en forme matricielle de jeux de séquences basée sur une localisation des caractères peut être de taille rédhibitoire. Nous proposons donc de localiser des patrons plutôt que des caractères. Nous présentons deux méthodes ad-hoc, l’une basée sur un parcours de treillis et l’autre sur la programmation dynamique. / Recent works have shown the relevance of constraint programming to tackle data mining tasks. This thesis follows this approach and addresses motif discovery in sequential data. We focus in particular, in the case of classified sequences, on the search for motifs that best fit each individual class. We propose a language of constraints over matrix domains to model such problems. The language assumes a preprocessing of the data set (e.g., by pre-computing the locations of each character in each sequence) and views a motif as the choice of a sub-matrix (i.e., characters, sequences, and locations). We introduce different matrix constraints (compatibility of locations with the database, class covering, location-based character ordering common to sequences, etc.) and address two NP-complete problems: the search for class-specific totally ordered motifs (e.g., exclusive subsequences) or partially ordered motifs. We provide two CSP models that rely on global constraints to prove exclusivity. We then present a memetic algorithm that uses this CSP model during initialisation and intensification. This hybrid approach proves competitive compared to the pure CSP approach as shown by experiments carried out on protein sequences. Lastly, we investigate data set preprocessing based on patterns rather than characters, in order to reduce the size of the resulting matrix domain. To this end, we present and compare two alternative methods, one based on lattice search, the other on dynamic programming.

Page generated in 0.0537 seconds