|
Showing results 18721 to 18740 of 96342
 Contributions théorique et numérique à l'étude de la photoréfractivité dans les semi-conducteurs II-VINguyen, Ngoc Duy  Master's dissertation (1998) Detailed reference viewed: 24 (5 ULg) Contributions to Batch Mode Reinforcement LearningFonteneau, Raphaël Doctoral thesis (2011) This dissertation presents various research contributions published during these four years of PhD in the field of batch mode reinforcement learning, which studies optimal control problems for which the ... [more ▼] This dissertation presents various research contributions published during these four years of PhD in the field of batch mode reinforcement learning, which studies optimal control problems for which the only information available on the system dynamics and the reward function is gathered in a set of trajectories. We first focus on deterministic problems in continuous spaces. In such a context, and under some assumptions related to the smoothness of the environment, we propose a new approach for inferring bounds on the performance of control policies. We also derive from these bounds a new inference algorithm for generalizing the information contained in the batch collection of trajectories in a cautious manner. This inference algorithm as itself lead us to propose a min max generalization framework. When working on batch mode reinforcement learning problems, one has also often to consider the problem of generating informative trajectories. This dissertation proposes two different approaches for addressing this problem. The first approach uses the bounds mentioned above to generate data tightening these bounds. The second approach proposes to generate data that are predicted to generate a change in the inferred optimal control policy. While the above mentioned contributions consider a deterministic framework, we also report on two research contributions which consider a stochastic setting. The first one addresses the problem of evaluating the expected return of control policies in the presence of disturbances. The second one proposes a technique for selecting relevant variables in a batch mode reinforcement learning context, in order to compute simplified control policies that are based on smaller sets of state variables. [less ▲] Detailed reference viewed: 78 (15 ULg) Contributions to Constitutive Las of Metals: Micro-Macro and Damage Models (Part C and D)Habraken, Anne Thèse d’agrégation de l’enseignement supérieur (2001) Detailed reference viewed: 14 (3 ULg)Contributions to Constitutive Laws of Metals: Micro-Macro and Damage Models (Part A & B)Habraken, Anne Thèse d’agrégation de l’enseignement supérieur (2001) Detailed reference viewed: 44 (11 ULg)Contributions to decision tree induction: bias/variance tradeoff and time series classificationGeurts, Pierre Doctoral thesis (2002) Because of the rapid progress of computer and information technology, large amounts of data are nowadays available in a lot of domains. Automatic learning aims at developing algorithms able to produce ... [more ▼] Because of the rapid progress of computer and information technology, large amounts of data are nowadays available in a lot of domains. Automatic learning aims at developing algorithms able to produce synthetic high-level information, or models, from this data. Learning algorithms are generally evaluated according to three different criteria: interpretability (how well the model helps to understand the data), predictive accuracy (how well the model can predict unseen situations), and computational efficiency (how fast is the algorithm and how it scales to large databases). This thesis explores two issues in automatic learning: the improvement of the well-known decision tree induction method and the problem of learning classification models for time series data. Decision tree induction method is an automatic learning algorithm which focuses on the modeling of input/output relationships. While this algorithm is among the fastest and most interpretable methods, its accuracy is not always competitive with respect to other algorithms. It is commonly admitted that this suboptimality is due to the excessive variance of this method. We first carry out an empirical study which shows quantitatively how important this variance is, i.e. how strongly decision trees depend on the random nature of the database used to infer them. These experiments confirm that this variance is detrimental not only from the point of view of accuracy but also from the point of view of interpretability. With the goal of improving both interpretability and accuracy, we consider three variance reduction techniques for decision trees. First, in the goal of improving mainly interpretability, we propose several methods which try to stabilize the parameters chosen during tree induction. While these methods succeed in reducing the variability of the parameters, they produce only a slight improvement of the accuracy. Then, we consider perturb and combine algorithms (e.g. bagging, boosting) which consist in combining the predictions of several models obtained by randomizing in some way the learning process. Inspired by the high variance of the parameters defining a decision tree, we propose an extremely randomized decision tree induction algorithm, called extra-tree, which chooses all parameters at random during induction. The aggregation of several of these extra-trees gives an important reduction of variance and this algorithm compares favorably in terms of accuracy and computational efficiency with both bagging and boosting. Because of the randomization of the parameters, the resulting method is also competitive with classical decision tree induction in terms of computational efficiency. In addition to these two approaches, we propose a ``dual'' perturb and combine algorithm which delays the perturbation at the prediction stage and hence requires only one model. In combination with decision tree, this method actually bridges the gap between single decision trees and perturb and combine algorithms. Of the first, it saves the interpretability (by using only one model), and with perturb and combine algorithm, it shares some of the accuracy (by reducing the variance). The second topic of the thesis is the problem of time series classification. The most direct way to solve this problem is to apply existing learning algorithms on low-level variables which correspond to the values of a time series at several time points. Experiments with the tree-based algorithms studied in the first part of the thesis shows that this approach is limited. A variance reduction techniques is then proposed specifically for this kind of data which consists in aggregating the prediction given by a classification model for subsequences of time series. Since this method does not provide interpretable models, we propose a second method which extends decision tree tests by allowing them to detect local shift invariant properties, or patterns, in time series. The study proposed in this part of the thesis is only a first step in the domain but our conclusions give some future work directions for handling complex type of data with automatic learning methods. [less ▲] Detailed reference viewed: 171 (16 ULg)  Contributions to early land plant research (XIV IBC).Gerrienne, Philippe ; in Review of Palaeobotany and Palynology (2001), 116 Detailed reference viewed: 3 (0 ULg)Contributions to recent advances in the field of steel joints. Column bases and further configurations for beam-to-column joints and column basesJaspart, Jean-Pierre Thèse d’agrégation de l’enseignement supérieur (1997) Detailed reference viewed: 144 (3 ULg)Contributions to Recognizability: Self-generating Sets, Decidability, Automaticity and Multidimensional SetsLacroix, Anne Doctoral thesis (2013) In this thesis, we study and answer several questions concerning recognizability of integer sets by finite automata. Each particular problem is the focus of a chapter. First, we study the recognizability ... [more ▼] In this thesis, we study and answer several questions concerning recognizability of integer sets by finite automata. Each particular problem is the focus of a chapter. First, we study the recognizability of the so-called self-generating sets, initially introduced by C. Kimberling. In the second part, we study the syntactic complexity of any ultimately periodic set and we use our results to give an alternative decision procedure for a well-known decidability problem. Next, we give bounds on the automaticity of three different languages: the language of primitive words over a finite alphabet, the language of unbordered words over a finite alphabet and the language of representations of monic irreducible polynomials over a finite fields. Finally, we characterize the multidimensional sets that are recognizable in all abstract numeration systems. [less ▲] Detailed reference viewed: 25 (11 ULg)Contributions to the development of a test-day model adjusted to New Zealand dairy cattleVanderick, Sylvie Master's dissertation (2004) A test-day model was developed in order to study the feasibility and the usefulness of a random regression test-day model based on dairy cattle data from New Zealand. (Co)variance components were ... [more ▼] A test-day model was developed in order to study the feasibility and the usefulness of a random regression test-day model based on dairy cattle data from New Zealand. (Co)variance components were estimated first within breeds then across breeds for Holstein-Friesian and Jersey animals. Results from single breed analysis showed that there are several differences between the two main breeds of dairy cattle in New Zealand. This was confirmed with the genetic correlations below one (r²= 0.926 for milk in first lactation) for breed specific additive genetic effects provided by the across breeds analysis. The specificity of the model used was first the capacity for taking account of these differences between breeds since more than one third of the dairy population are crossbred; and secondly the estimation of different breeding values considering each animal as a potential mate within each of the two breeds. From genetic evaluations, ranking changes in sires were observed by using the model with breed specific additive genetic effects and a model assuming the same breed composition for all animals. This first investigation showed the opportunity to carry on more research in the way to develop an accurate test-day model permitting to evaluate the crossbred dairy cattle of New Zealand and to perform an optimal selection of animals in function of the mating strategy to apply. [less ▲] Detailed reference viewed: 33 (7 ULg)Contributions to the knowledge about the seismic behaviour of typical Belgian masonry housesDegée, Hervé ; Plumier, André in Proceedings of the 8th International Masonry Conference (2010) Detailed reference viewed: 52 (4 ULg)Contributions to the modelling of fires in buildings and their effects on structuresFranssen, Jean-Marc Scientific conference (1998, July 24) Detailed reference viewed: 3 (1 ULg)Il contributo della ricerca: i primi dati e le prospettive per la pratica; ; Miscioscia, Marina Conference (2011, February 25) Detailed reference viewed: 9 (4 ULg)Contributor to Small Animal Ophthalmology, 4th Edition - A Problem-Oriented Approach (DVD)Monclin, Sébastien in R.L. Peiffer; S.M. Petersen-Jones (Eds.) Small Animal Ophthalmology, 4th Edition - A Problem-Oriented Approach (2009) Detailed reference viewed: 9 (1 ULg)Contro la pastasciutta : Marinetti’s Futurist LunchDelville, Michel in Intervalles (2007), 2 Detailed reference viewed: 22 (2 ULg)Control and simulation of doubly fed induction generator for variable speed wind turbine systems based on an integrated finite element approachChen, Qiongzhong ; ; Bruls, Olivier in Proceedings of the European Wind Energy Conference and Exhibition 2011 (EWEA 2011) (2011, March) Regarding renewable energy and environment- friendly issues, wind energy nowadays has become the fastest-growing energy source in the world, and thus attracts a lot of research interest in the wind ... [more ▼] Regarding renewable energy and environment- friendly issues, wind energy nowadays has become the fastest-growing energy source in the world, and thus attracts a lot of research interest in the wind turbine generation system. A doubly fed induction generator (DFIG) is used for variable speed operation in a wind turbine system to extract more power. Following a systematic approach, this paper investigates on the modelling and simulation of wind turbine generating systems using the flexible multibody simulation software SAMCEF/MECANO [1]. The objective of this work is to analyze the control-generator-structure interactions in a wind turbine system. Firstly, an extension of the finite element method is integrated into the flexible multibody dynamics solver, and thus extends the solver to represent mechatronic systems in a strongly-coupled way. Secondly, DFIG and the control systems are modularly modeled for the wind turbine package. Control of DFIG for grid synchronization and power optimization are elaborated in detail, and the methods are validated through a 2MW DFIG wind turbine prototype model. At the end, a systematic system model of wind turbine structure connected with the DFIG generating system is presented, which provides the dynamic analysis for the whole system in an overall range. [less ▲] Detailed reference viewed: 56 (11 ULg)Control and Voting Power in Complex Shareholding NetworksCrama, Yves ; Leruth, Luc ; Conference (2011, June) Detailed reference viewed: 29 (3 ULg)Control and voting power in corporate networks: Concepts and computational aspectsCrama, Yves ; Leruth, Luc in European Journal of Operational Research (2007), 178(3), 879-893 This paper proposes to rely on power indices to measure the amount of control held by individual shareholders in corporate networks. The value of the indices is determined by a complex voting game viewed ... [more ▼] This paper proposes to rely on power indices to measure the amount of control held by individual shareholders in corporate networks. The value of the indices is determined by a complex voting game viewed as the composition of interlocked weighted majority games; the compound game reflects the structure of shareholdings. The paper describes an integrated algorithmic approach which allows to deal efficiently with the complexity of computing power indices in shareholding networks, irrespective of their size or structure. In particular, the approach explicitly accounts for the presence of float and of cyclic shareholding relationships. It has been successfully applied to the analysis of real-world financial networks. (c) 2006 Elsevier B.V. All rights reserved. [less ▲] Detailed reference viewed: 46 (5 ULg)Control and Voting Power in Shareholding NetworksCrama, Yves Conference (2010, March 18) Detailed reference viewed: 3 (1 ULg)Control charts in a delivery chain frameworkFaraz, Alireza ; ; Heuchenne, Cédric et alE-print/Working paper (n.d.) We call attention to delivery chain systems as an area for statistical process control research and propose using elapsed time variables as delivery-quality measures. We propose a Shewhart-style control ... [more ▼] We call attention to delivery chain systems as an area for statistical process control research and propose using elapsed time variables as delivery-quality measures. We propose a Shewhart-style control charting strategy that we term ‘bundling’ which describes creating and averaging related time variables to allow for a reduction in the number of charts needed for monitoring delivery chain systems. The major assumption is that several related delivery chain system time-variables (DCSTV) representing origin to destination pairs that deliver products to a common geographical region would shift together when being monitored for delay. In the context of strategic monitoring of aggregate systems, bundled charts can effectively signal when origin-destination pairs shift in a particular direction. Given that some of the DCSTV can be ‘bundled’ for representation on Shewhart charts, decision makers can be provided an accurate assessment of the predictability (or unpredictability) of a delivery chain. Providing more accurate promises of delivery time through the use of Shewhart-style charts will ultimately reduce levels of safety stock which results in a reduction in inventory holding costs. [less ▲] Detailed reference viewed: 15 (0 ULg) |
||
