Examinando por Materia "algorithms"
Mostrando 1 - 5 de 5
Resultados por página
Opciones de ordenación
Ítem Adaptive LAMDA applied to identify and regulate a process with variable dead time(Institute of Electrical and Electronics Engineers Inc., 2020-01-01) Morales L.; Pozo D.; Aguilar J.; Rosales A.; Universidad EAFIT. Departamento de Ingeniería de Sistemas; I+D+I en Tecnologías de la Información y las ComunicacionesIn this paper, an adaptive intelligent controller based on the fuzzy algorithm called LAMDA (Learning Algorithm for Multivariable Data Analysis) is presented in order to identify and regulate a process with variable dead time. The original algorithm has been used for supervised and unsupervised learning, whose main field of application is the identification of functional states of the systems. In this work a modification of LAMDA has been implemented which is capable of online learning using hybrid techniques. The proposal consists of two stages: training stage to learn about the unknown plant in order to establish initial parameters to the controller, and a second phase, called application, in which the control strategy is updated using online learning. The proposed method is tested in the control objective of regulation of a process with variable dead time, to analyze the viability of its utilization in these types of systems in which their dynamics are variable and unknown. © 2020 IEEE.Ítem CME – A web application framework learning technique based on concerns, micro-learning and examples(SPRINGER, 2018-01-01) Correa, D.; Isaza, F.A.; Mazo, R.; Giraldo, G.L.; Universidad EAFIT. Departamento de Ingeniería de Sistemas; I+D+I en Tecnologías de la Información y las ComunicacionesNowadays the use of Web Application Frameworks has been the default choice to develop software applications inside the web domain. These frameworks provide many benefits such as: faster development, enhanced security, and clearer structure of code. However, web application frameworks have a high learning curve and each time someone wants to learn a framework, a recurrent question arises: How to reduce that learning curve and be able to use a new web application framework in a short time? Several learning techniques have been used in the context of frameworks; for instance, example-based learning, tutorials, cookbooks, and tools. Nevertheless, those approaches have their own limitations and the question about what is the best way to learn soon those frameworks remains open. In this paper, we introduce a new web application framework learning technique called CME (Concerns, Micro-learning and Examples). This technique defines a customized learning path that a novice developer should follow to learn to use a web application framework. The learning path was built as a mix of concerns, micro-learning and example-based learning concepts. To validate this technique, we carry out a quasi-experiment about a web application framework learning and the development of small web applications with: (i) the use of CME and (ii) the use of a cookbook. The quasi-experiment results showed preliminary evidence that when using CME, novice developers become more operational to develop small web applications in a reduced period of time compared with the use of a cookbook. © Springer International Publishing AG, part of Springer Nature 2018.Ítem Multiobjective model for multicast overlay networks over IP/MPLS using MOEA(2008-01-01) Montoya, J.; Donoso, Y.; Montoya, E.; Echeverri, D.; Universidad EAFIT. Departamento de Ingeniería de Sistemas; I+D+I en Tecnologías de la Información y las ComunicacionesMulticast plays an important role in supporting a new generation of applications. At present and for different reasons, technical and non-technical, multicast IP hasn't yet been totally adopted for Internet. During recent years, an active area of research is that of implementing this kind of traffic in the application layer where the multicast functionality isnt a responsibility of the routers but that of the hosts, which we know as Multicast Overlay Networks (MON). In this article, routing in an MON is put forward as a multiobjective optimization problem (MOP) where two functions are optimized: 1) the total end to end delay of the multicast tree and 2) the maximum link utilization. The simultaneous optimization of these two functions is an NP-Complete problem and to solve this we suggest using Multiobjective Evolutionary Algorithms (MOEA), specifically NSGA-II.Ítem Observations about an approximate algorithm for the point robot motion planning problem(IEEE Computer Society, 2002-01-01) Trefftz, C.; Trefftz, H.; Universidad EAFIT. Departamento de Ingeniería de Sistemas; I+D+I en Tecnologías de la Información y las ComunicacionesObservations about an approximate parallel algorithm for the point robot motion planning problem are presented. The algorithm solves not only the original problem, but related problems as well. © 2002 IEEE.Ítem Parallelizing AES on multicores and GPUs(2011-01-01) Ortega, J.; Trefftz, H.; Trefftz, C.; Universidad EAFIT. Departamento de Ingeniería de Sistemas; I+D+I en Tecnologías de la Información y las ComunicacionesThe AES block cipher cryptographic algorithm is widely used and it is resource intensive. An existing sequential open source implementation of the algorithm was parallelized on multi-core microprocessors and GPUs. Performance results are presented. © 2011 IEEE.