http://hdl.handle.net/2117/3486 Tue, 21 May 2013 22:08:24 GMT 2013-05-21T22:08:24Z webmaster.bupc@upc.edu Universitat Politècnica de Catalunya. Servei de Biblioteques i Documentació no http://hdl.handle.net/2117/19025 Title: The parameters of Menzerath-Altmann law in genomes Authors: Baixeries i Juvillà, Jaume; Hernández Fernández, Antonio; Forns, Núria; Ferrer Cancho, Ramon Abstract: The relationship between the size of the whole and the size of the parts in language and music is known to follow the Menzerath-Altmann law at many levels of description (morphemes, words, sentences, …). Qualitatively, the law states that the larger the whole, the smaller its parts, e.g. the longer a word (in syllables) the shorter its syllables (in letters or phonemes). This patterning has also been found in genomes: the longer a genome (in chromosomes), the shorter its chromosomes (in base pairs). However, it has been argued recently that mean chromosome length is trivially a pure power function of chromosome number with an exponent of -1. The functional dependency between mean chromosome size and chromosome number in groups of organisms from three different kingdoms is studied. The fit of a pure power function yields exponents between -1.6 and 0.1. It is shown that an exponent of -1 is unlikely for fungi, gymnosperm plants, insects, reptiles, ray-finned fishes and amphibians. Even when the exponent is very close to -1, adding an exponential component is able to yield a better fit with regard to a pure power-law in plants, mammals, ray-finned fishes and amphibians. The parameters of the Menzerath-Altmann law in genomes deviate significantly from a power law with a -1 exponent with the exception of birds and cartilaginous fishes. Fri, 26 Apr 2013 18:45:28 GMT http://hdl.handle.net/2117/19025 2013-04-26T18:45:28Z Baixeries i Juvillà, Jaume; Hernández Fernández, Antonio; Forns, Núria; Ferrer Cancho, Ramon no The relationship between the size of the whole and the size of the parts in language and music is known to follow the Menzerath-Altmann law at many levels of description (morphemes, words, sentences, …). Qualitatively, the law states that the larger the whole, the smaller its parts, e.g. the longer a word (in syllables) the shorter its syllables (in letters or phonemes). This patterning has also been found in genomes: the longer a genome (in chromosomes), the shorter its chromosomes (in base pairs). However, it has been argued recently that mean chromosome length is trivially a pure power function of chromosome number with an exponent of -1. The functional dependency between mean chromosome size and chromosome number in groups of organisms from three different kingdoms is studied. The fit of a pure power function yields exponents between -1.6 and 0.1. It is shown that an exponent of -1 is unlikely for fungi, gymnosperm plants, insects, reptiles, ray-finned fishes and amphibians. Even when the exponent is very close to -1, adding an exponential component is able to yield a better fit with regard to a pure power-law in plants, mammals, ray-finned fishes and amphibians. The parameters of the Menzerath-Altmann law in genomes deviate significantly from a power law with a -1 exponent with the exception of birds and cartilaginous fishes. http://hdl.handle.net/2117/18572 Title: Learning theory through videos: a teaching experience in a theoretical course based on self-learning videos and problem-solving sessions Authors: Arias Vicente, Marta; Creus López, Carles; Gascón Caro, Adrià; Godoy Balil, Guillem Abstract: In this paper we describe a teaching experience applied to a theoretical course tought in a computer science degree. The main feature of our experiment is the introduction of videos specifically designed for self-learning as part of the learning process. Master classes are replaced by working sessions in which the involvement of students gains prominence. The teacher explains almost nothing in class. Instead, most of the time is devoted to the presentation of solutions to exercises assigned to students in advance. All presentations are done by students, and the teacher only intervenes in order to complete explanations and correct mistakes. The result of our experiment is promising from several perspectives. The exam results are better with the new approach. The students learn to learn on their own and take better advantage of the time in class. The work load is uniformly distributed along the course. The new approach also benefits the teacher since he/she spends considerably less time preparing theory lectures, and gets continuous feedback to better follow the students developement. The videos are valuable in themselves and have been made publicly available. In fact, our students prefer them to a master class. They can pause, rewind and replay the video, take a rest, and postpone the lecture if necessary. Moreover, the interest for these videos goes beyond our university boundaries: according to the visits procedence and posted comments, they are being used by students from other countries. Wed, 03 Apr 2013 11:30:11 GMT http://hdl.handle.net/2117/18572 2013-04-03T11:30:11Z Arias Vicente, Marta; Creus López, Carles; Gascón Caro, Adrià; Godoy Balil, Guillem no Self-learning, e-Learning, Videos, Teaching, Theory In this paper we describe a teaching experience applied to a theoretical course tought in a computer science degree. The main feature of our experiment is the introduction of videos specifically designed for self-learning as part of the learning process. Master classes are replaced by working sessions in which the involvement of students gains prominence. The teacher explains almost nothing in class. Instead, most of the time is devoted to the presentation of solutions to exercises assigned to students in advance. All presentations are done by students, and the teacher only intervenes in order to complete explanations and correct mistakes. The result of our experiment is promising from several perspectives. The exam results are better with the new approach. The students learn to learn on their own and take better advantage of the time in class. The work load is uniformly distributed along the course. The new approach also benefits the teacher since he/she spends considerably less time preparing theory lectures, and gets continuous feedback to better follow the students developement. The videos are valuable in themselves and have been made publicly available. In fact, our students prefer them to a master class. They can pause, rewind and replay the video, take a rest, and postpone the lecture if necessary. Moreover, the interest for these videos goes beyond our university boundaries: according to the visits procedence and posted comments, they are being used by students from other countries. http://hdl.handle.net/2117/18260 Title: Learning probabilistic automata : a study in state distinguishability Authors: Balle Pigem, Borja de; Castro Rabal, Jorge; Gavaldà Mestre, Ricard Abstract: Known algorithms for learning PDFA can only be shown to run in time polynomial in the so-called distinguishability μ of the target machine, besides the number of states and the usual accuracy and confidence parameters. We show that the dependence on μ is necessary in the worst case for every algorithm whose structure resembles existing ones. As a technical tool, a new variant of Statistical Queries termed View the MathML source-queries is defined. We show how to simulate View the MathML source-queries using classical Statistical Queries and show that known PAC algorithms for learning PDFA are in fact statistical query algorithms. Our results include a lower bound: every algorithm to learn PDFA with queries using a reasonable tolerance must make Ω(1/μ1−c) queries for every c>0. Finally, an adaptive algorithm that PAC-learns w.r.t. another measure of complexity is described. This yields better efficiency in many cases, while retaining the same inevitable worst-case behavior. Our algorithm requires fewer input parameters than previously existing ones, and has a better sample bound. Wed, 13 Mar 2013 13:49:57 GMT http://hdl.handle.net/2117/18260 2013-03-13T13:49:57Z Balle Pigem, Borja de; Castro Rabal, Jorge; Gavaldà Mestre, Ricard no Known algorithms for learning PDFA can only be shown to run in time polynomial in the so-called distinguishability μ of the target machine, besides the number of states and the usual accuracy and confidence parameters. We show that the dependence on μ is necessary in the worst case for every algorithm whose structure resembles existing ones. As a technical tool, a new variant of Statistical Queries termed View the MathML source-queries is defined. We show how to simulate View the MathML source-queries using classical Statistical Queries and show that known PAC algorithms for learning PDFA are in fact statistical query algorithms. Our results include a lower bound: every algorithm to learn PDFA with queries using a reasonable tolerance must make Ω(1/μ1−c) queries for every c>0. Finally, an adaptive algorithm that PAC-learns w.r.t. another measure of complexity is described. This yields better efficiency in many cases, while retaining the same inevitable worst-case behavior. Our algorithm requires fewer input parameters than previously existing ones, and has a better sample bound. http://hdl.handle.net/2117/18232 Title: A graphical tool for describing the temporal evolution of clusters in financial stock markets Authors: Arratia Quesada, Argimiro Alejandro; Cabaña, Ana Alejandra Tue, 12 Mar 2013 16:27:48 GMT http://hdl.handle.net/2117/18232 2013-03-12T16:27:48Z Arratia Quesada, Argimiro Alejandro; Cabaña, Ana Alejandra no http://hdl.handle.net/2117/17754 Title: Spectral learning of general weighted automata via constrained matrix completion Authors: Balle Pigem, Borja de; Mohri, Mehryar Abstract: Many tasks in text and speech processing and computational biology require estimating functions mapping strings to real numbers. A broad class of such functions can be defined by weighted automata. Spectral methods based on the singular value decomposition of a Hankel matrix have been recently proposed for learning a probability distribution represented by a weighted automaton from a training sample drawn according to this same target distribution. In this paper, we show how spectral methods can be extended to the problem of learning a general weighted automaton from a sample generated by an arbitrary distribution. The main obstruction to this approach is that, in general, some entries of the Hankel matrix may be missing. We present a solution to this problem based on solving a constrained matrix completion problem. Combining these two ingredients, matrix completion and spectral method, a whole new family of algorithms for learning general weighted automata is obtained. We present generalization bounds for a particular algorithm in this family. The proofs rely on a joint stability analysis of matrix completion and spectral learning. Thu, 14 Feb 2013 10:29:29 GMT http://hdl.handle.net/2117/17754 2013-02-14T10:29:29Z Balle Pigem, Borja de; Mohri, Mehryar no Many tasks in text and speech processing and computational biology require estimating functions mapping strings to real numbers. A broad class of such functions can be defined by weighted automata. Spectral methods based on the singular value decomposition of a Hankel matrix have been recently proposed for learning a probability distribution represented by a weighted automaton from a training sample drawn according to this same target distribution. In this paper, we show how spectral methods can be extended to the problem of learning a general weighted automaton from a sample generated by an arbitrary distribution. The main obstruction to this approach is that, in general, some entries of the Hankel matrix may be missing. We present a solution to this problem based on solving a constrained matrix completion problem. Combining these two ingredients, matrix completion and spectral method, a whole new family of algorithms for learning general weighted automata is obtained. We present generalization bounds for a particular algorithm in this family. The proofs rely on a joint stability analysis of matrix completion and spectral learning. http://hdl.handle.net/2117/17639 Title: Border algorithms for computing Hasse diagrams of arbitrary lattices Authors: Balcázar Navarro, José Luis; Tirnauca, Cristina Abstract: Lattices are mathematical structures with many applications in computer science; among these, we are interested in fields like data mining, machine learning, or knowledge discovery in databases. One well-established use of lattice theory is in formal concept analysis (FCA), where the concept lattice with its diagram graph allows the visualization and summarization of data in a more concise representation. In the Data Mining community, the same mathematical notions (often under additional “frequency” constraints that bound from below the size of the support set) are studied under the banner of Closed-Set Mining. In these applications, each dataset consists of transactions, also called objects, each of which, besides having received a unique identifier, consists of a set of items or attributes taken from a previously agreed finite set. A concept is a pair formed by a set of transactions —the extent set or support set of the concept— and a set of attributes —the intent set of the concept— defined as the set of all those attributes that are shared by all the transactions present in the extent. Some data analysis processes are based on the family of all intents (the “closures” stemming from the dataset); but others require to determine also their order relation, which is a finite lattice, in the form of a line graph (the Hasse diagram). Mon, 11 Feb 2013 15:24:34 GMT http://hdl.handle.net/2117/17639 2013-02-11T15:24:34Z Balcázar Navarro, José Luis; Tirnauca, Cristina no Lattices, Hasse diagrams, Border algorithms Lattices are mathematical structures with many applications in computer science; among these, we are interested in fields like data mining, machine learning, or knowledge discovery in databases. One well-established use of lattice theory is in formal concept analysis (FCA), where the concept lattice with its diagram graph allows the visualization and summarization of data in a more concise representation. In the Data Mining community, the same mathematical notions (often under additional “frequency” constraints that bound from below the size of the support set) are studied under the banner of Closed-Set Mining. In these applications, each dataset consists of transactions, also called objects, each of which, besides having received a unique identifier, consists of a set of items or attributes taken from a previously agreed finite set. A concept is a pair formed by a set of transactions —the extent set or support set of the concept— and a set of attributes —the intent set of the concept— defined as the set of all those attributes that are shared by all the transactions present in the extent. Some data analysis processes are based on the family of all intents (the “closures” stemming from the dataset); but others require to determine also their order relation, which is a finite lattice, in the form of a line graph (the Hasse diagram). http://hdl.handle.net/2117/17434 Title: Bootstrapping and learning PDFA in data streams Authors: Balle Pigem, Borja de; Castro Rabal, Jorge; Gavaldà Mestre, Ricard Abstract: Markovian models with hidden state are widely-used formalisms for modeling sequential phenomena. Learnability of these models has been well studied when the sample is given in batch mode, and algorithms with PAC-like learning guarantees exist for specic classes of models such as Probabilistic Deterministic Finite Automata (PDFA). Here we focus on PDFA and give an algorithm for infering models in this class under the stringent data stream scenario: unlike existing methods, our algorithm works incrementally and in one pass, uses memory sublinear in the stream length, and processes input items in amortized constant time. We provide rigorous PAC-like bounds for all of the above, as well as an evaluation on synthetic data showing that the algorithm performs well in practice. Our algorithm makes a key usage of several old and new sketching techniques. In particular, we develop a new sketch for implementing bootstrapping in a streaming setting which may be of independent interest. In experiments we have observed that this sketch yields important reductions in the examples required for performing some crucial statistical tests in our algorithm. Mon, 21 Jan 2013 09:08:53 GMT http://hdl.handle.net/2117/17434 2013-01-21T09:08:53Z Balle Pigem, Borja de; Castro Rabal, Jorge; Gavaldà Mestre, Ricard no Markovian models with hidden state are widely-used formalisms for modeling sequential phenomena. Learnability of these models has been well studied when the sample is given in batch mode, and algorithms with PAC-like learning guarantees exist for specic classes of models such as Probabilistic Deterministic Finite Automata (PDFA). Here we focus on PDFA and give an algorithm for infering models in this class under the stringent data stream scenario: unlike existing methods, our algorithm works incrementally and in one pass, uses memory sublinear in the stream length, and processes input items in amortized constant time. We provide rigorous PAC-like bounds for all of the above, as well as an evaluation on synthetic data showing that the algorithm performs well in practice. Our algorithm makes a key usage of several old and new sketching techniques. In particular, we develop a new sketch for implementing bootstrapping in a streaming setting which may be of independent interest. In experiments we have observed that this sketch yields important reductions in the examples required for performing some crucial statistical tests in our algorithm. http://hdl.handle.net/2117/17082 Title: The exponent of Zipf’s law in language ontogeny. Authors: Baixeries i Juvillà, Jaume; Ferrer Cancho, Ramon Fri, 07 Dec 2012 11:07:01 GMT http://hdl.handle.net/2117/17082 2012-12-07T11:07:01Z Baixeries i Juvillà, Jaume; Ferrer Cancho, Ramon no http://hdl.handle.net/2117/17062 Title: Computing Functional Dependencies with Pattern Structures Authors: Baixeries i Juvillà, Jaume; Kaytoue, Mehdi; Napoli, Amedeo Abstract: The treatment of many-valued data with FCA has been achieved by means of scaling. This method has some drawbacks, since the size of the resulting formal contexts depends usually on the number of di erent values that are present in a table, which can be very large. Pattern structures have been proved to deal with many-valued data, offering a viable and sound alternative to scaling in order to represent and analyze sets of many-valued data with FCA. Functional dependencies have already been dealt with FCA using the binarization of a table, that is, creating a formal context out of a set of data. Unfortunately, although this method is standard and simple, it has an important drawback, which is the fact that the resulting context is quadratic in number of objects w.r.t. the original set of data. In this paper, we examine how we can extract the functional dependencies that hold in a set of data using pattern structures. This allows to build an equivalent concept lattice avoiding the step of binarization, and thus comes with better concept representation and computation. Mon, 03 Dec 2012 16:19:30 GMT http://hdl.handle.net/2117/17062 2012-12-03T16:19:30Z Baixeries i Juvillà, Jaume; Kaytoue, Mehdi; Napoli, Amedeo no The treatment of many-valued data with FCA has been achieved by means of scaling. This method has some drawbacks, since the size of the resulting formal contexts depends usually on the number of di erent values that are present in a table, which can be very large. Pattern structures have been proved to deal with many-valued data, offering a viable and sound alternative to scaling in order to represent and analyze sets of many-valued data with FCA. Functional dependencies have already been dealt with FCA using the binarization of a table, that is, creating a formal context out of a set of data. Unfortunately, although this method is standard and simple, it has an important drawback, which is the fact that the resulting context is quadratic in number of objects w.r.t. the original set of data. In this paper, we examine how we can extract the functional dependencies that hold in a set of data using pattern structures. This allows to build an equivalent concept lattice avoiding the step of binarization, and thus comes with better concept representation and computation. http://hdl.handle.net/2117/17023 Title: A logic programming approach to parsing and production in fluid construction grammar Authors: Sierra Santibáñez, Josefina Abstract: This paper presents a Logic Programming approach to parsing and production in Fluid Construction Grammar (FCG). It builds on previous work on the formalisation of FCG in terms of First Order Logic (FOL) concepts, more specifically on the definition of its core inference operations, unification and merge, in terms of FOL unification and search in the space of a particular set of FOL terms called structure arrangements. An implementation of such inference operations based on Logic Programming and Artificial Intelligence techniques such as unification and heuristic search is outlined. Mon, 26 Nov 2012 15:45:23 GMT http://hdl.handle.net/2117/17023 2012-11-26T15:45:23Z Sierra Santibáñez, Josefina no Logic programming, Parsing, Fluid construction grammar This paper presents a Logic Programming approach to parsing and production in Fluid Construction Grammar (FCG). It builds on previous work on the formalisation of FCG in terms of First Order Logic (FOL) concepts, more specifically on the definition of its core inference operations, unification and merge, in terms of FOL unification and search in the space of a particular set of FOL terms called structure arrangements. An implementation of such inference operations based on Logic Programming and Artificial Intelligence techniques such as unification and heuristic search is outlined. http://hdl.handle.net/2117/16934 Title: Two measures of objective novelty in association rule mining Authors: Balcázar Navarro, José Luis Abstract: Association rule mining is well-known to depend heavily on a support threshold parameter, and on one or more thresholds for intensity of implication; among these measures, confidence is most often used and, sometimes, related alternatives such as lift, leverage, improvement, or all-confidence are employed, either separately or jointly with confidence. We remain within the support-and-confidence framework in an attempt at studying complementary notions, which have the goal of measuring relative forms of objective novelty or surprisingness of each individual rule with respect to other rules that hold in the same dataset. We measure novelty through the extent to which the confidence value is robust, taken relative to the confidences of related (for instance, logically stronger) rules, as opposed to the absolute consideration of the single rule at hand. We consider two variants of this idea and analyze their logical and algorithmic properties. Since this approach has the drawback of requiring further parameters, we also propose a framework in which the user sets a single parameter, of quite clear intuitive semantics, from which the corresponding thresholds for confidence and novelty are computed. Fri, 16 Nov 2012 10:11:03 GMT http://hdl.handle.net/2117/16934 2012-11-16T10:11:03Z Balcázar Navarro, José Luis no Association rule mining is well-known to depend heavily on a support threshold parameter, and on one or more thresholds for intensity of implication; among these measures, confidence is most often used and, sometimes, related alternatives such as lift, leverage, improvement, or all-confidence are employed, either separately or jointly with confidence. We remain within the support-and-confidence framework in an attempt at studying complementary notions, which have the goal of measuring relative forms of objective novelty or surprisingness of each individual rule with respect to other rules that hold in the same dataset. We measure novelty through the extent to which the confidence value is robust, taken relative to the confidences of related (for instance, logically stronger) rules, as opposed to the absolute consideration of the single rule at hand. We consider two variants of this idea and analyze their logical and algorithmic properties. Since this approach has the drawback of requiring further parameters, we also propose a framework in which the user sets a single parameter, of quite clear intuitive semantics, from which the corresponding thresholds for confidence and novelty are computed. http://hdl.handle.net/2117/16933 Title: Mining educational data for patterns with negations and high confidence boost Authors: Balcázar Navarro, José Luis; Tirnauca, Cristina; Zorrilla Pantaleón, Marta Elena Abstract: Association rules constitute a well-known and widely employed data mining technique. We study their applicability in Educational Data Mining. We develop a case study of datasets from that eld: logs of an e-learning platform. We demonstrate that it is convenient to analyze such datasets in terms of association rules that relate not only presence of items in each of the transactions, but also their absence. To cope with the algorithmic di culties and the large output, we apply a new heuristic regarding the support of negative attributes, complementing two previously studied contributions: a basis for closure-oriented notions of redundancy and a notion of novelty called the con dence boost. Our ndings have been validated through interactions with end-user experts, namely, the instructors in whose virtual learning courses the datasets had their origin. Thu, 15 Nov 2012 12:40:19 GMT http://hdl.handle.net/2117/16933 2012-11-15T12:40:19Z Balcázar Navarro, José Luis; Tirnauca, Cristina; Zorrilla Pantaleón, Marta Elena no Association rules constitute a well-known and widely employed data mining technique. We study their applicability in Educational Data Mining. We develop a case study of datasets from that eld: logs of an e-learning platform. We demonstrate that it is convenient to analyze such datasets in terms of association rules that relate not only presence of items in each of the transactions, but also their absence. To cope with the algorithmic di culties and the large output, we apply a new heuristic regarding the support of negative attributes, complementing two previously studied contributions: a basis for closure-oriented notions of redundancy and a notion of novelty called the con dence boost. Our ndings have been validated through interactions with end-user experts, namely, the instructors in whose virtual learning courses the datasets had their origin. http://hdl.handle.net/2117/16913 Title: Objective novelty of association rules: measuring the confidence boost Authors: Balcázar Navarro, José Luis Abstract: On sait bien que la confiance des régles d’association n’est pas vraiment satisfaisant comme mésure d’interêt. Nous proposons, au lieu de la substituer par des autres mésures (soit, en l’employant de façon conjointe a des autres mésures), évaluer la nouveauté de chaque régle par comparaison de sa confiance par rapport á des régles plus fortes qu’on trouve au même ensemble de données. C’est á dire, on considère un seuil “relative” de confiance au lieu du seuil absolute habituel. Cette idée se précise avec la magnitude du “confidence boost”, mésurant l’increment rélative de confiance prés des régles plus fortes. Nous prouvons que nôtre proposte peut remplacer la “confidence width” et le blockage de régles employés a des publications précedentes. Tue, 13 Nov 2012 13:28:37 GMT http://hdl.handle.net/2117/16913 2012-11-13T13:28:37Z Balcázar Navarro, José Luis no On sait bien que la confiance des régles d’association n’est pas vraiment satisfaisant comme mésure d’interêt. Nous proposons, au lieu de la substituer par des autres mésures (soit, en l’employant de façon conjointe a des autres mésures), évaluer la nouveauté de chaque régle par comparaison de sa confiance par rapport á des régles plus fortes qu’on trouve au même ensemble de données. C’est á dire, on considère un seuil “relative” de confiance au lieu du seuil absolute habituel. Cette idée se précise avec la magnitude du “confidence boost”, mésurant l’increment rélative de confiance prés des régles plus fortes. Nous prouvons que nôtre proposte peut remplacer la “confidence width” et le blockage de régles employés a des publications précedentes. http://hdl.handle.net/2117/16907 Title: Closed-set-based discovery of representative association rules revisited Authors: Balcázar Navarro, José Luis; Tirnauca, Cristina Abstract: The output of an association rule miner is often huge in practice. This is why several concise lossless representations have been proposed, such as the “essential” or “representative” rules. We revisit the algorithm given by Kryszkiewicz (Int. Symp. Intelligent Data Analysis 2001, Springer-Verlag LNCS 2189, 350–359) for mining representative rules. We show that its output is sometimes incomplete, due to an oversight in its mathematical validation, and we propose an alternative complete generator that works within only slightly larger running times. Tue, 13 Nov 2012 11:34:41 GMT http://hdl.handle.net/2117/16907 2012-11-13T11:34:41Z Balcázar Navarro, José Luis; Tirnauca, Cristina no The output of an association rule miner is often huge in practice. This is why several concise lossless representations have been proposed, such as the “essential” or “representative” rules. We revisit the algorithm given by Kryszkiewicz (Int. Symp. Intelligent Data Analysis 2001, Springer-Verlag LNCS 2189, 350–359) for mining representative rules. We show that its output is sometimes incomplete, due to an oversight in its mathematical validation, and we propose an alternative complete generator that works within only slightly larger running times. http://hdl.handle.net/2117/16067 Title: Energy-efficient and multifaceted resource management for profit-driven virtualized data centers Authors: Goiri Presa, Íñigo; Berral García, Josep Lluís; Fitó, Josep Oriol; Julià Massó, Ferran; Nou Castell, Ramon; Guitart Fernández, Jordi; Gavaldà Mestre, Ricard; Torres Viñals, Jordi Abstract: As long as virtualization has been introduced in data centers, it has been opening new chances for resource management. Nowadays, it is not just used as a tool for consolidating underused nodes and save power; it also allows new solutions to well-known challenges, such as heterogeneity management. Virtualization helps to encapsulate Web-based applications or HPC jobs in virtual machines (VMs) and see them as a single entity which can be managed in an easier and more efficient way. We propose a new scheduling policy that models and manages a virtualized data center. It focuses on the allocation of VMs in data center nodes according to multiple facets to optimize the provider’s profit. In particular, it considers energy efficiency, virtualization overheads, and SLA violation penalties, and supports the outsourcing to external providers. The proposed approach is compared to other common scheduling policies, demonstrating that a provider can improve its benefit by 30% and save power while handling other challenges, such as resource outsourcing, in a better and more intuitive way than other typical approaches do. Sat, 16 Jun 2012 10:58:35 GMT http://hdl.handle.net/2117/16067 2012-06-16T10:58:35Z Goiri Presa, Íñigo; Berral García, Josep Lluís; Fitó, Josep Oriol; Julià Massó, Ferran; Nou Castell, Ramon; Guitart Fernández, Jordi; Gavaldà Mestre, Ricard; Torres Viñals, Jordi no As long as virtualization has been introduced in data centers, it has been opening new chances for resource management. Nowadays, it is not just used as a tool for consolidating underused nodes and save power; it also allows new solutions to well-known challenges, such as heterogeneity management. Virtualization helps to encapsulate Web-based applications or HPC jobs in virtual machines (VMs) and see them as a single entity which can be managed in an easier and more efficient way. We propose a new scheduling policy that models and manages a virtualized data center. It focuses on the allocation of VMs in data center nodes according to multiple facets to optimize the provider’s profit. In particular, it considers energy efficiency, virtualization overheads, and SLA violation penalties, and supports the outsourcing to external providers. The proposed approach is compared to other common scheduling policies, demonstrating that a provider can improve its benefit by 30% and save power while handling other challenges, such as resource outsourcing, in a better and more intuitive way than other typical approaches do.