ISEL - Matemática - Comunicações
Permanent URI for this collection
Browse
Browsing ISEL - Matemática - Comunicações by Issue Date
Now showing 1 - 10 of 41
Results Per Page
Sort Options
- Symbolic Knowledge Extraction from Trained Neural Networks Governed by Lukasiewicz LogicsPublication . Leandro, Carlos; Pinheiro Pita, Helder Jorge; Monteiro, LuísThis work describes a methodology to extract symbolic rules from trained neural networks. In our approach, patterns on the network are codified using formulas on a Lukasiewicz logic. For this we take advantage of the fact that every connective in this multi-valued logic can be evaluated by a neuron in an artificial network having, by activation function the identity truncated to zero and one. This fact simplifies symbolic rule extraction and allows the easy injection of formulas into a network architecture. We trained this type of neural network using a back-propagation algorithm based on Levenderg-Marquardt algorithm, where in each learning iteration, we restricted the knowledge dissemination in the network structure. This makes the descriptive power of produced neural networks similar to the descriptive power of Lukasiewicz logic language, minimizing the information loss on the translation between connectionist and symbolic structures. To avoid redundance on the generated network, the method simplifies them in a pruning phase, using the "Optimal Brain Surgeon" algorithm. We tested this method on the task of finding the formula used on the generation of a given truth table. For real data tests, we selected the Mushrooms data set, available on the UCI Machine Learning Repository.
- Generalized models from Beta(p,2) densities with strong allee effect: dynamical approachPublication . Aleixo, Sandra; Rocha, J. LeonelA dynamical approach to study the behaviour of generalized populational growth models from Bets(p, 2) densities, with strong Allee effect, is presented. The dynamical analysis of the respective unimodal maps is performed using symbolic dynamics techniques. The complexity of the correspondent discrete dynamical systems is measured in terms of topological entropy. Different populational dynamics regimes are obtained when the intrinsic growth rates are modified: extinction, bistability, chaotic semistability and essential extinction.
- Damping optimization of viscoelastic laminated sandwich structures using the direct multisearch methodPublication . J.F.A. Madeira; Araújo, A. L.; Soares, C. M. Mota; Soares, C. A. MotaA multiobjective approach for optimization of passive damping for vibration reduction in sandwich structures is presented in this paper. A layerwise finite element model for sandwich plates with viscoelastic core and anisotropic laminated face layers is used along with the complex modulus approach and the dynamic problem is solved in the frequency domain. Constrained optimization is conducted for maximisation of modal loss factors and minimisation of weight of sandwich beams and plates with elastic laminated constraining layers and a viscoelastic core, with layer thickness and laminate layer ply orientation angles as design variables. The problem is solved using the Direct MultiSearch (DMS) solver for derivative-free multiobjective optimization and solutions are compared with alternative ones obtained using genetic algorithms. DMS is a solver for multiobjective optimization problems which does not use any derivatives of the objective functions. It is based on a novel technique called direct multisearch, developed by extending direct search from single to multiobjective optimisation.
- Joining models with stair nestingPublication . Fernandes, Célia; Ramos, PauloIn the stair nested designs with u factors we have u steps and a(1), ... , a(u) "active" levels. We would have a(1) observations with different levels for the first factor each of them nesting a single level of each of the remaining factors; next a(2) observations with level a(1) + 1 for the first factor and distinct levels for the second factor each nesting a fixed level of each of the remaining factors, and so on. So the number of level combinations is Sigma(u)(i=1) a(i). In meta-analysis joint treatment of different experiments is considered. Joining the corresponding models may be useful to carry out that analysis. In this work we want joining L models with stair nesting.
- Study of the interactions in a three-way crossed classification modelPublication . Ramos, Paulo; Fernandes, CéliaCrossed classification models are applied in many investigations taking in consideration the existence of interaction between all factors or, in alternative, excluding all interactions, and in this case only the effects and the error term are considered. In this work we use commutative Jordan algebras in the study of the algebraic structure of these designs and we use them to obtain similar designs where only some of the interactions are considered. We finish presenting the expressions of the variance componentes estimators.
- Modeling Allee Effect from Beta(p, 2) DensitiesPublication . Rocha, J. Leonel; Aleixo, SandraIn this work we develop and investigate generalized populational growth models, adjusted from Beta(p, 2) densities, with Allee effect. The use of a positive parameter leads the presented generalization, which yields some more flexible models with variable extinction rates. An Allee limit is incorporated so that the models under study have strong Allee effect.
- Numerical 3D modeling of heat transfer in human tissues for microwave radiometry monitoring of Brown fat metabolismoPublication . Rodrigues, Dário B.; Maccarini, Paolo F.; Salahi, Sara; Colebeck, Erin; Topsakal, Erdem; Pereira, Pedro Jorge da Silva; Limão-Vieira, Paulo; Stauffer, Paul R.Background: Brown adipose tissue (BAT) plays an important role in whole body metabolism and could potentially mediate weight gain and insulin sensitivity. Although some imaging techniques allow BAT detection, there are currently no viable methods for continuous acquisition of BAT energy expenditure. We present a non-invasive technique for long term monitoring of BAT metabolism using microwave radiometry. Methods: A multilayer 3D computational model was created in HFSS™ with 1.5 mm skin, 3-10 mm subcutaneous fat, 200 mm muscle and a BAT region (2-6 cm3) located between fat and muscle. Based on this model, a log-spiral antenna was designed and optimized to maximize reception of thermal emissions from the target (BAT). The power absorption patterns calculated in HFSS™ were combined with simulated thermal distributions computed in COMSOL® to predict radiometric signal measured from an ultra-low-noise microwave radiometer. The power received by the antenna was characterized as a function of different levels of BAT metabolism under cold and noradrenergic stimulation. Results: The optimized frequency band was 1.5-2.2 GHz, with averaged antenna efficiency of 19%. The simulated power received by the radiometric antenna increased 2-9 mdBm (noradrenergic stimulus) and 4-15 mdBm (cold stimulus) corresponding to increased 15-fold BAT metabolism. Conclusions: Results demonstrated the ability to detect thermal radiation from small volumes (2-6 cm3) of BAT located up to 12 mm deep and to monitor small changes (0.5°C) in BAT metabolism. As such, the developed miniature radiometric antenna sensor appears suitable for non-invasive long term monitoring of BAT metabolism.
- Electricity spot prices structural changes in the Iberian electricity marketPublication . Bolas, João; Sousa, Jorge A. M.; Martins, Ana Alexandra; Lagarto, JoãoIn recent years, the power sector has undergone a restructuring process in many economies in the world. This movement towards liberalization led to the establishment of electricity markets that promote the competitiveness of the production and trading segments of the power sector. In these markets, the agents have to deal with frequent electricity price changes leading to different strategies in their daily bidding behavior. There are a set of variables that can have an impact in the electricity price definition, such as: fuel prices, CO 2 emissions prices, electricity production and demand. This paper proposes to analyze structural changes in the Iberian electricity market price between two periods of time: 2007/2008 and 2010/2011. For this purpose, three quantitative analysis methods were used: correlation, causality and Principal Components. Results suggest that the electricity price had a structural change between the analyzed periods, in particular the increasing importance of special regime production.
- Optimal renewable generation mix of hydro, wind and photovoltaic for integration into the Portuguese power systemPublication . Sousa, Jorge A. M.; Martins, Ana AlexandraIn spite of the advantages of renewable energy sources (RES), their variability and uncertainty raises important issues in power systems operation, such as the need to balance demand with a highly variable and uncertain power generation. In this context, is of utmost importance the choice of the optimal renewable generation mix to be integrated in the power system in order to accomplish the established renewable energy targets with the lowest technical impacts. With this aim, this paper presents a methodology that supports decision making on the renewable energy policy by deriving the optimal renewable generation mix from different available technologies - hydro, wind and photovoltaic - that integrates a given amount of electricity from renewable sources, taking into account the variability of the renewable generation mix and the target share of renewable generation. This methodology is applied to a case study of one month using renewable generation data of the Portuguese power system.
- Modeling the strategic behavior of the iberian electricity market producers using time series analysisPublication . Faria, Ricardo; Sousa, Jorge A. M.; Martins, Ana Alexandra; Lagarto, JoãoThe Iberian Electricity Market (MIBEL) emerges in the context of the integration and cooperation between the Portuguese and Spanish electricity markets, in response to the European Union incentive for regional electricity markets creation. The present study, focus on the modeling and forecasting of the hourly competitive strategies of the electricity producers in the MIBEL. For this analysis, the studied variable was the MIBEL's conjectural variation, which estimates the level of competitiveness of the electricity producers on the day-ahead electricity market. The methodology adopted for forecasting was time series analysis, using ARIMA and exponential smoothing models. The results obtained show that the estimated models that best suit the hourly MIBEL conjectural variation forecast were mainly of the ARIMA seasonal type with daily seasonality, followed by ARIMA non-seasonal type models. It was also observed, that the selected models were mainly estimated with a time series of 5 working days.