Optimization methods known from the literature include gradient techniques and evolutionary algorithms. The main idea of gradient methods is to calculate the gradient of the objective function at the actual point and ...
详细信息
Optimization methods known from the literature include gradient techniques and evolutionary algorithms. The main idea of gradient methods is to calculate the gradient of the objective function at the actual point and then to step towards better values according to this value. Evolutionary algorithms imitate a simplified abstract model of evolution observed in nature. Memetic algorithms traditionally combine evolutionary and gradient techniques to exploit the advantages of both methods. Our current research aims to discover the properties, especially the efficiency (i.e. the speed of convergence) of particular evolutionary and memetic algorithms. For this purpose the techniques are compared by applying them on several numerical optimization benchmark functions and on fuzzy rule base identification.
Unlike standard output error models, both input and output observations of errors-in-variables systems are corrupted with noise. As practical applications often fall into the errors-in-variables category, estimation m...
详细信息
Unlike standard output error models, both input and output observations of errors-in-variables systems are corrupted with noise. As practical applications often fall into the errors-in-variables category, estimation methods that can simultaneously derive model and noise parameters are of particular interest. In this paper, we explore separation mechanisms applied on the combined input and output observation matrix to partition observations into two sets, after which it is possible to perform parameter estimation over the individual sets, and the estimates may, in turn, be compared. Minimizing the distance between parameter estimates, it is shown that we may infer a noise structure. Once a noise structure estimate is at hand, a maximum likelihood estimation may yield model parameter estimates.
The performance of information systems is an important consideration. With the help of proper performance models and evaluation algorithms, performance metrics can be predicted accurately. The goal of our work is to i...
详细信息
Model-driven approaches try to improve the productivity of the traditional software development by focusing on a special problem using high abstraction models. Comparing to the source code, specific visualization of m...
详细信息
Model-driven approaches try to improve the productivity of the traditional software development by focusing on a special problem using high abstraction models. Comparing to the source code, specific visualization of models inspired by these domains is very advantageous and closer to the human thinking. Though, in many cases the details can be efficiently edited only in textual form: visual models are sometimes not expressive or effective enough, and could be circumstantial to use. The common problem of typical modeling tools is that they provide support for only one kind of model representation, but it would be more effective if two or more representation possibilities could be mixed. In this paper, we suggest an approach, which supports the concurrent evolution of the artifacts in both textual and visual way. The declarative textual syntax mapping is defined in a way to facilitate bi-directional conversion. To synchronize, and merge the changes of the different kinds of artifacts incrementally, a three-way comparison is applied.
Model transformation is a crucial aspect of model-driven software development. With the help of model transformation, we can generate source code or other models from software models. However, a recurring problem in s...
详细信息
Model transformation is a crucial aspect of model-driven software development. With the help of model transformation, we can generate source code or other models from software models. However, a recurring problem in software development is the fact that source and target models coexist and they evolve independently. This means that after the initial transformation phase the target models are often changed by the developers and the target model will no longer be consistent with the source model. Therefore, the models need to be synchronized. Model transformation is one option to implement the synchronization process. However, it is very complex because the transformation has to recognize which elements should be updated and which elements should be left unaltered during synchronization. In addition, a clear backward transformation usually does not exist from the target to the source domain. In this paper, we present an algorithm that facilitates incremental model synchronization in model transformation systems. The feasibility of the approach is also illustrated by a real world case study; we synchronize user interfaces of different mobile platforms.
In the recent years, software modeling became an essential part of designing large software systems. However, efficient model processing techniques are required, unless we use models only for documentation purposes. O...
详细信息
In the recent years, software modeling became an essential part of designing large software systems. However, efficient model processing techniques are required, unless we use models only for documentation purposes. One way to improve efficiency is to apply model transformations in parallel. This paper introduces a truly parallel model transformation framework focusing on a new pattern matching approach. By using this new approach, the framework is now scalable and much more efficient than any parallel model transformation approach before.
When building topic based document classifiers, feature selection is a key step: features not holding any information about the topic of a document introduce only unnecessary noise during the classification. In a dist...
详细信息
When building topic based document classifiers, feature selection is a key step: features not holding any information about the topic of a document introduce only unnecessary noise during the classification. In a distributed environment, when the nodes are interacting, the locally retrieved features and the their attributes must be shared to have at every node a more accurate estimation of the global classifier. When expanding the knowledge of the local classifiers, to reduce costs, the network traffic should be kept to a minimum. We propose a probabilistic model for a keyword selection method which makes a more thorough analysis possible and can be used as a baseline when sharing information in a distributed environment. It can be used for incrementally building up the distributed classifiers ensuring minimal network traffic. This model can be refined later on by sending more content-related information to achieve higher performance. This probabilistic model together with experimental results are presented in this paper.
Localization is a fundamental task in mobile robotics and in indoor environments we can use various sensors to solve this problem. In the Intelligent Space environment we can use laser range finders or ultrasonic posi...
详细信息
Localization is a fundamental task in mobile robotics and in indoor environments we can use various sensors to solve this problem. In the Intelligent Space environment we can use laser range finders or ultrasonic positioning systems to localize and track mobile robots. Nevertheless, our final goal is to substitute these sensors and accomplish this task using just cameras. In this paper, we show the feasibility of determining the robot's location based on the images received from a single camera. In our experimental room we used a surveillance camera, which can be controlled to pan/tilt in order to change the point of view. The camera had been mounted on the ceiling and six preset positions were selected to cover the whole area. The object recognition is based on colour space filtering and contour detection. Finally, the contours of the detected objects are transformed from the image space to the world coordinate system and the polygons are reduced to simpler ones. The system is able to detect not only the position of the objects, but their orientations.
Sustainable waste management systems necessarily include environmental, economic, social, institutional, legal and technical aspects. As a result of the incompleteness and multiple uncertainties occurring in sustainab...
详细信息
The aim of this paper is to present a control scheme for controlling the temperature in a slab which is not directly measurable. The process can only be controlled by manipulating the ambient temperature (boundary con...
详细信息
The aim of this paper is to present a control scheme for controlling the temperature in a slab which is not directly measurable. The process can only be controlled by manipulating the ambient temperature (boundary control). We describe the physical process and derive its (distributed) transfer function (DTF). The sensitivity of the model for one of the parameters is analyzed. We describe a model reduction technique which provides an accurate approximation of the distributed transfer function. Based on the reduced order model we present an Internal Model Control (IMC) scheme applied to the heat process and analyze the *** order model
暂无评论