1982, vol. 6, núm. 2http://hdl.handle.net/2099/38202024-03-28T21:25:11Z2024-03-28T21:25:11ZÍndexQüestiióhttp://hdl.handle.net/2099/54992020-07-21T18:58:36Z2008-05-29T16:21:18ZÍndex
Qüestiió
2008-05-29T16:21:18ZQüestiióEstimation of random survival function: a linear approachQuesada Paloma, VicenteGarcía Pérez, Alfonsohttp://hdl.handle.net/2099/45532015-08-03T01:04:50Z2008-03-10T13:33:17ZEstimation of random survival function: a linear approach
Quesada Paloma, Vicente; García Pérez, Alfonso
In the first part of this work, a Survival function is considered which is supposed to be an Exponential Gamma Process. The main statistical and probability properties of this process and its Bayesian interpretation are considered.
In the second part, the problem to estimate, from a Bayesian view point, the Survival function is considered, looking for the Bayes rule inside of the set of linear combinations of a given set of sample functions.
We finish with an estimation, in the same situation like before, of the survival mean time, and the i-th moment about the origin of the Survival function.
2008-03-10T13:33:17ZQuesada Paloma, VicenteGarcía Pérez, AlfonsoIn the first part of this work, a Survival function is considered which is supposed to be an Exponential Gamma Process. The main statistical and probability properties of this process and its Bayesian interpretation are considered.
In the second part, the problem to estimate, from a Bayesian view point, the Survival function is considered, looking for the Bayes rule inside of the set of linear combinations of a given set of sample functions.
We finish with an estimation, in the same situation like before, of the survival mean time, and the i-th moment about the origin of the Survival function.Reduction des machines séquentielles non déterministesZahnd, Jacqueshttp://hdl.handle.net/2099/45522015-08-03T01:04:51Z2008-03-10T13:27:11ZReduction des machines séquentielles non déterministes
Zahnd, Jacques
En partant de la notion classique de machine séquentielle incomplètement spécifiée, on introduit une généralisation de ce modèle, sous la forme de tables d'états qui associent à chaque état présent et à chaque signal d'entrée un ensemble d'états futurs et un ensemble de signaux de sortie. On définit ensuite une fonction de réponse qui caractérise le comportement entrée-sortie d'une telle machine, ce qui permet de poser en termes exacts le problème de sa réduction. On passe en revue quelques applications possibles de ce modèle: synthese des systèmes digitaux, réseaux de Petri, microprogrammes non déterministes. Ensuite on aborde le problème de la réduction de ces machines. On généralise la notion classique de recouvrement compatible et l'on montre qu'elle permet de réduire une machine non déterministe. On termine par l'énoncé des principaux problèmes ouverts qui peuvent faire l'objet d'une suite de ce travail.; Starting from the classical notion of an incompletely specified sequential machine, a generalisation of this model is introduced in the form of state tables which associate with each present state/input combination a set of next states and a set of outputs. The response function, characterizing the input/output behaviour of such a machine is then defined, which makes possible to state in rigorous terms the problem of its minimization. Some possible applications of the model are surveyed: design of digital systems, some types of Petri nets, non deterministic microprograms. The minimization problem is then approached. The classical notion of a compatible cover is suitably generalized, and is shown to supply a possible solution. The paper ends with the statement of some open problems.
2008-03-10T13:27:11ZZahnd, JacquesEn partant de la notion classique de machine séquentielle incomplètement spécifiée, on introduit une généralisation de ce modèle, sous la forme de tables d'états qui associent à chaque état présent et à chaque signal d'entrée un ensemble d'états futurs et un ensemble de signaux de sortie. On définit ensuite une fonction de réponse qui caractérise le comportement entrée-sortie d'une telle machine, ce qui permet de poser en termes exacts le problème de sa réduction. On passe en revue quelques applications possibles de ce modèle: synthese des systèmes digitaux, réseaux de Petri, microprogrammes non déterministes. Ensuite on aborde le problème de la réduction de ces machines. On généralise la notion classique de recouvrement compatible et l'on montre qu'elle permet de réduire une machine non déterministe. On termine par l'énoncé des principaux problèmes ouverts qui peuvent faire l'objet d'une suite de ce travail.
Starting from the classical notion of an incompletely specified sequential machine, a generalisation of this model is introduced in the form of state tables which associate with each present state/input combination a set of next states and a set of outputs. The response function, characterizing the input/output behaviour of such a machine is then defined, which makes possible to state in rigorous terms the problem of its minimization. Some possible applications of the model are surveyed: design of digital systems, some types of Petri nets, non deterministic microprograms. The minimization problem is then approached. The classical notion of a compatible cover is suitably generalized, and is shown to supply a possible solution. The paper ends with the statement of some open problems.Zero or near-to-zero Lagrange multipliers in linearly constrained nonlinear programmingEscudero, L. F.http://hdl.handle.net/2099/45512015-08-03T01:04:53Z2008-03-10T13:24:27ZZero or near-to-zero Lagrange multipliers in linearly constrained nonlinear programming
Escudero, L. F.
We discuss in this work the using of Lagrange multipliers estimates in linearly constrained nonlinear programming algorithms and the implication of zero or near-to-zero Lagrange multipliers. Some methods for estimating the tendency of the multipliers are proposed in the context of a given algorithm.
2008-03-10T13:24:27ZEscudero, L. F.We discuss in this work the using of Lagrange multipliers estimates in linearly constrained nonlinear programming algorithms and the implication of zero or near-to-zero Lagrange multipliers. Some methods for estimating the tendency of the multipliers are proposed in the context of a given algorithm.Un algoritmo heurístico lagrangiano para el problema de localización de plantas con capacidadesBarceló Bugeda, JaimeCasanovas Garcia, Josephttp://hdl.handle.net/2099/45502021-04-21T10:09:12Z2008-03-10T13:21:49ZUn algoritmo heurístico lagrangiano para el problema de localización de plantas con capacidades
Barceló Bugeda, Jaime; Casanovas Garcia, Josep
Las técnicas lagrangianas se han aplicado con frecuencia al problema de localización de plantas cuando no intervienen las capacidades, y en algunos casos han demostrado su utilidad incluso cuando se tienen en cuenta restricciones adicionales. Nuestro trabajo estudia la aplicación de estas técnicas al problema de localización de plantas cuando intervienen las capacidades, en el caso particular en que el modelo considerado es entero puro. Se han tenido en cuenta varias descomposiciones lagrangianas, y para alguna de ellas se han diseñado algoritmos heurísticos para resolver los subproblemas lagrangianos.
Las heurísticas consisten en un procedimiento con dos fases. En la primera (fase de localización) se define un conjunto de multiplicadores a partir del análisis del dual de la relajación LP, y se efectúa una selección de emplazamientos para las plantas, mientras que la segunda (fase de afectación) asigna los centros a las plantas considerando el subproblema resultante como un caso particular del problema generalizado de asignación. Varias heurísticas han sido estudiadas en esta segunda fase, basadas en la descomposición en una colección de subproblemas de tipo knapsack mediante la definición de un conjunto de penalizaciones, o en el análisis del duality gap intentando reducirlo. Se incluyen los resultados de las experiencias realizadas.; Lagrangean techniques have been widely applied to the uncapacitated plant location problem, and in some cases they have been proven to be succesfull even when additional constraints are taken into account. We study the aplication of these techniques to the capacitated plant location problem when the model considered is a pure integer one. Several lagrangean decompositions have been studied, and for some of them heuristic algorithms have been designed to solve the lagrangean subproblems, the heuristics sonsisting of a two phase procedure. The first (location phase) defines a set of multipliers from the analysis of the dual LP relaxation, and makes a choice of the plants to be located, while the second (allocation phase) assigns the centers to the plants considering the resulting subproblem as a particular case of the general assignment problem. Several heuristics have been studied for this second phase based either on a decomposition on knapsack type subproblems through the definition of a set of penalties, or on looking into the duality gap trying to reduce it. Computational experience is reported.
2008-03-10T13:21:49ZBarceló Bugeda, JaimeCasanovas Garcia, JosepLas técnicas lagrangianas se han aplicado con frecuencia al problema de localización de plantas cuando no intervienen las capacidades, y en algunos casos han demostrado su utilidad incluso cuando se tienen en cuenta restricciones adicionales. Nuestro trabajo estudia la aplicación de estas técnicas al problema de localización de plantas cuando intervienen las capacidades, en el caso particular en que el modelo considerado es entero puro. Se han tenido en cuenta varias descomposiciones lagrangianas, y para alguna de ellas se han diseñado algoritmos heurísticos para resolver los subproblemas lagrangianos.
Las heurísticas consisten en un procedimiento con dos fases. En la primera (fase de localización) se define un conjunto de multiplicadores a partir del análisis del dual de la relajación LP, y se efectúa una selección de emplazamientos para las plantas, mientras que la segunda (fase de afectación) asigna los centros a las plantas considerando el subproblema resultante como un caso particular del problema generalizado de asignación. Varias heurísticas han sido estudiadas en esta segunda fase, basadas en la descomposición en una colección de subproblemas de tipo knapsack mediante la definición de un conjunto de penalizaciones, o en el análisis del duality gap intentando reducirlo. Se incluyen los resultados de las experiencias realizadas.
Lagrangean techniques have been widely applied to the uncapacitated plant location problem, and in some cases they have been proven to be succesfull even when additional constraints are taken into account. We study the aplication of these techniques to the capacitated plant location problem when the model considered is a pure integer one. Several lagrangean decompositions have been studied, and for some of them heuristic algorithms have been designed to solve the lagrangean subproblems, the heuristics sonsisting of a two phase procedure. The first (location phase) defines a set of multipliers from the analysis of the dual LP relaxation, and makes a choice of the plants to be located, while the second (allocation phase) assigns the centers to the plants considering the resulting subproblem as a particular case of the general assignment problem. Several heuristics have been studied for this second phase based either on a decomposition on knapsack type subproblems through the definition of a set of penalties, or on looking into the duality gap trying to reduce it. Computational experience is reported.