Optimization of Weighted Finite State Transducer for Speech Recognition

There is considerable interest in creating embedded, speech recognition hardware using the weighted finite state transducer (WFST) technique but there are performance and memory usage challenges. Two system optimization techniques are presented to address this; one approach improves token propagation by removing the WFST epsilon input arcs; another one-pass, adaptive pruning algorithm gives a dramatic reduction in active nodes to be computed. Results for memory and bandwidth are given for a 5,000 word vocabulary giving a better practical performance than conventional WFST; this is then exploited in an adaptive pruning algorithm that reduces the active nodes from 30,000 down to 4,000 with only a 2 percent sacrifice in speech recognition accuracy; these optimizations lead to a more simplified design with deterministic performance.

DFIG Machine Design for Maximizing Power Output Based on Surrogate Optimization Algorithm

This paper presents a surrogate-model based optimization of a doubly-fed induction generator (DFIG) machine winding design for maximizing power yield. Based on site-specific wind profile data and the machine’s previous operational performance, the DFIG’s stator and rotor windings are optimized to match the maximum efficiency with operating conditions for rewinding purposes. The particle swarm optimization (PSO)-based surrogate optimization techniques are used in conjunction with the finite element method (FEM) to optimize the machine design utilizing the limited available information for the site-specific wind profile and generator operating conditions. A response surface method in the surrogate model is developed to formulate the design objectives and constraints. Besides, the machine tests and efficiency calculations follow IEEE standard 112-B. Numerical and experimental results validate the effectiveness of the proposed technologies.

Analysis and Optimization of Machining Process Using Evolutionary Algorithms

To ensure quality of machined products at minimum machining costs and maximum machining effectiveness, it is very important to select optimum parameters when metal cutting machine tools are employed. Traditionally, the experience of the operator plays a major role in the selection of optimum metal cutting conditions. However, attaining optimum values each time by even a skilled operator is difficult. The non-linear nature of the machining process has compelled engineers to search for more effective methods to attain optimization. The design objective preceding most engineering design activities is simply to minimize the cost of production or to maximize the production efficiency. The main aim of research work reported here is to build robust optimization algorithms by exploiting ideas that nature has to offer from its backyard and using it to solve real world optimization problems in manufacturing processes.In this thesis, after conducting an exhaustive literature review, several optimization techniques used in various manufacturing processes have been identified. The selection of optimal cutting parameters, like depth of cut, feed and speed is a very important issue for every machining process. Experiments have been designed using Taguchi technique and dry turning of SS420 has been performed on Kirlosker turn master 35 lathe. Analysis using S/N and ANOVA were performed to find the optimum level and percentage of contribution of each parameter. By using S/N analysis the optimum machining parameters from the experimentation is obtained.Optimization algorithms begin with one or more design solutions supplied by the user and then iteratively check new design solutions, relative search spaces in order to achieve the true optimum solution. A mathematical model has been developed using response surface analysis for surface roughness and the model was validated using published results from literature.Methodologies in optimization such as Simulated annealing (SA), Particle Swarm Optimization (PSO), Conventional Genetic Algorithm (CGA) and Improved Genetic Algorithm (IGA) are applied to optimize machining parameters while dry turning of SS420 material. All the above algorithms were tested for their efficiency, robustness and accuracy and observe how they often outperform conventional optimization method applied to difficult real world problems. The SA, PSO, CGA and IGA codes were developed using MATLAB. For each evolutionary algorithmic method, optimum cutting conditions are provided to achieve better surface finish.The computational results using SA clearly demonstrated that the proposed solution procedure is quite capable in solving such complicated problems effectively and efficiently. Particle Swarm Optimization (PSO) is a relatively recent heuristic search method whose mechanics are inspired by the swarming or collaborative behavior of biological populations. From the results it has been observed that PSO provides better results and also more computationally efficient.Based on the results obtained using CGA and IGA for the optimization of machining process, the proposed IGA provides better results than the conventional GA. The improved genetic algorithm incorporating a stochastic crossover technique and an artificial initial population scheme is developed to provide a faster search mechanism. Finally, a comparison among these algorithms were made for the specific example of dry turning of SS 420 material and arriving at optimum machining parameters of feed, cutting speed, depth of cut and tool nose radius for minimum surface roughness as the criterion. To summarize, the research work fills in conspicuous gaps between research prototypes and industry requirements, by simulating evolutionary procedures seen in nature that optimize its own systems.

Adaptive Real-time Anomaly-based Intrusion Detection using Data Mining and Machine Learning Techniques

Die zunehmende Vernetzung der Informations- und Kommunikationssysteme führt zu einer weiteren Erhöhung der Komplexität und damit auch zu einer weiteren Zunahme von Sicherheitslücken. Klassische Schutzmechanismen wie Firewall-Systeme und Anti-Malware-Lösungen bieten schon lange keinen Schutz mehr vor Eindringversuchen in IT-Infrastrukturen. Als ein sehr wirkungsvolles Instrument zum Schutz gegenüber Cyber-Attacken haben sich hierbei die Intrusion Detection Systeme (IDS) etabliert. Solche Systeme sammeln und analysieren Informationen von Netzwerkkomponenten und Rechnern, um ungewöhnliches Verhalten und Sicherheitsverletzungen automatisiert festzustellen. Während signatur-basierte Ansätze nur bereits bekannte Angriffsmuster detektieren können, sind anomalie-basierte IDS auch in der Lage, neue bisher unbekannte Angriffe (Zero-Day-Attacks) frühzeitig zu erkennen. Das Kernproblem von Intrusion Detection Systeme besteht jedoch in der optimalen Verarbeitung der gewaltigen Netzdaten und der Entwicklung eines in Echtzeit arbeitenden adaptiven Erkennungsmodells. Um diese Herausforderungen lösen zu können, stellt diese Dissertation ein Framework bereit, das aus zwei Hauptteilen besteht. Der erste Teil, OptiFilter genannt, verwendet ein dynamisches "Queuing Concept", um die zahlreich anfallenden Netzdaten weiter zu verarbeiten, baut fortlaufend Netzverbindungen auf, und exportiert strukturierte Input-Daten für das IDS. Den zweiten Teil stellt ein adaptiver Klassifikator dar, der ein Klassifikator-Modell basierend auf "Enhanced Growing Hierarchical Self Organizing Map" (EGHSOM), ein Modell für Netzwerk Normalzustand (NNB) und ein "Update Model" umfasst. In dem OptiFilter werden Tcpdump und SNMP traps benutzt, um die Netzwerkpakete und Hostereignisse fortlaufend zu aggregieren. Diese aggregierten Netzwerkpackete und Hostereignisse werden weiter analysiert und in Verbindungsvektoren umgewandelt. Zur Verbesserung der Erkennungsrate des adaptiven Klassifikators wird das künstliche neuronale Netz GHSOM intensiv untersucht und wesentlich weiterentwickelt. In dieser Dissertation werden unterschiedliche Ansätze vorgeschlagen und diskutiert. So wird eine classification-confidence margin threshold definiert, um die unbekannten bösartigen Verbindungen aufzudecken, die Stabilität der Wachstumstopologie durch neuartige Ansätze für die Initialisierung der Gewichtvektoren und durch die Stärkung der Winner Neuronen erhöht, und ein selbst-adaptives Verfahren eingeführt, um das Modell ständig aktualisieren zu können. Darüber hinaus besteht die Hauptaufgabe des NNB-Modells in der weiteren Untersuchung der erkannten unbekannten Verbindungen von der EGHSOM und der Überprüfung, ob sie normal sind. Jedoch, ändern sich die Netzverkehrsdaten wegen des Concept drif Phänomens ständig, was in Echtzeit zur Erzeugung nicht stationärer Netzdaten führt. Dieses Phänomen wird von dem Update-Modell besser kontrolliert. Das EGHSOM-Modell kann die neuen Anomalien effektiv erkennen und das NNB-Model passt die Änderungen in Netzdaten optimal an. Bei den experimentellen Untersuchungen hat das Framework erfolgversprechende Ergebnisse gezeigt. Im ersten Experiment wurde das Framework in Offline-Betriebsmodus evaluiert. Der OptiFilter wurde mit offline-, synthetischen- und realistischen Daten ausgewertet. Der adaptive Klassifikator wurde mit dem 10-Fold Cross Validation Verfahren evaluiert, um dessen Genauigkeit abzuschätzen. Im zweiten Experiment wurde das Framework auf einer 1 bis 10 GB Netzwerkstrecke installiert und im Online-Betriebsmodus in Echtzeit ausgewertet. Der OptiFilter hat erfolgreich die gewaltige Menge von Netzdaten in die strukturierten Verbindungsvektoren umgewandelt und der adaptive Klassifikator hat sie präzise klassifiziert. Die Vergleichsstudie zwischen dem entwickelten Framework und anderen bekannten IDS-Ansätzen zeigt, dass der vorgeschlagene IDSFramework alle anderen Ansätze übertrifft. Dies lässt sich auf folgende Kernpunkte zurückführen: Bearbeitung der gesammelten Netzdaten, Erreichung der besten Performanz (wie die Gesamtgenauigkeit), Detektieren unbekannter Verbindungen und Entwicklung des in Echtzeit arbeitenden Erkennungsmodells von Eindringversuchen.

Simulation and optimisation techniques for sawmill yard operation : A literature review

Increasing costs and competitive business strategies are pushing sawmill enterprises to make an effort for optimization of their process management. Organizational decisions mainly concentrate on performance and reduction of operational costs in order to maintain profit margins. Although many efforts have been made, effective utilization of resources, optimal planning and maximum productivity in sawmill are still challenging to sawmill industries. Many researchers proposed the simulation models in combination with optimization techniques to address problems of integrated logistics optimization. The combination of simulation and optimization technique identifies the optimal strategy by simulating all complex behaviours of the system under consideration including objectives and constraints. During the past decade, an enormous number of studies were conducted to simulate operational inefficiencies in order to find optimal solutions. This paper gives a review on recent developments and challenges associated with simulation and optimization techniques. It was believed that the review would provide a perfect ground to the authors in pursuing further work in optimizing sawmill yard operations.

Artificial neural networks and clustering techniques applied in the reconfiguration of distribution systems

One objective of the feeder reconfiguration problem in distribution systems is to minimize the power losses for a specific load. For this problem, mathematical modeling is a nonlinear mixed integer problem that is generally hard to solve. This paper proposes an algorithm based on artificial neural network theory. In this context, clustering techniques to determine the best training set for a single neural network with generalization ability are also presented. The proposed methodology was employed for solving two electrical systems and presented good results. Moreover, the methodology can be employed for large-scale systems in real-time environment.

Short and long period optimization of drug doses in the treatment of AIDS

Técnicas de otimização numérica são úteis na solução de problemas de determinação da melhor entrada para sistemas descritos por modelos matemáticos e cujos objetivos podem ser expressos de uma maneira quantitativa. Este trabalho aborda o problema de otimizar as dosagens dos medicamentos no tratamento da AIDS em termos de um balanço entre a resposta terapêutica e os efeitos colaterais. Um modelo matemático para descrever a dinâmica do vírus HIV e células CD4 é utilizado para calcular a dosagem ótima do medicamento no tratamento a curto prazo de pacientes com AIDS por um método de otimização direta utilizando uma função custo do tipo Bolza. Os parâmetros do modelo foram ajustados com dados reais obtidos da literatura. Com o objetivo de simplificar os procedimentos numéricos, a lei de controle foi expressa em termos de uma expansão em séries que, após truncamento, permite obter controles sub-ótimos. Quando os pacientes atingem um estado clínico satisfatório, a técnica do Regulador Linear Quadrático (RLQ) é utilizada para determinar a dosagem permanente de longo período para os medicamentos. As dosagens calculadas utilizando a técnica RLQ , tendem a ser menores do que a equivalente terapia de dose constante em termos do expressivo aumento na contagem das células T+ CD4 e da redução da densidade de vírus livre durante um intervalo fixo de tempo.

A New Approach to Reducing Search Space and Increasing Efficiency in Simulation Optimization Problems via the Fuzzy-DEA-BCC

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

A comparison about evolutionary algorithms for optimum-path forest clustering optimization

In this paper we deal with the problem of boosting the Optimum-Path Forest (OPF) clustering approach using evolutionary-based optimization techniques. As the OPF classifier performs an exhaustive search to find out the size of sample's neighborhood that allows it to reach the minimum graph cut as a quality measure, we compared several optimization techniques that can obtain close graph cut values to the ones obtained by brute force. Experiments in two public datasets in the context of unsupervised network intrusion detection have showed the evolutionary optimization techniques can find suitable values for the neighborhood faster than the exhaustive search. Additionally, we have showed that it is not necessary to employ many agents for such task, since the neighborhood size is defined by discrete values, with constrain the set of possible solution to a few ones.

A Chaotic Approach of Differential Evolution Optimization Applied to Loudspeaker Design Problem

Over the past few years, the field of global optimization has been very active, producing different kinds of deterministic and stochastic algorithms for optimization in the continuous domain. These days, the use of evolutionary algorithms (EAs) to solve optimization problems is a common practice due to their competitive performance on complex search spaces. EAs are well known for their ability to deal with nonlinear and complex optimization problems. Differential evolution (DE) algorithms are a family of evolutionary optimization techniques that use a rather greedy and less stochastic approach to problem solving, when compared to classical evolutionary algorithms. The main idea is to construct, at each generation, for each element of the population a mutant vector, which is constructed through a specific mutation operation based on adding differences between randomly selected elements of the population to another element. Due to its simple implementation, minimum mathematical processing and good optimization capability, DE has attracted attention. This paper proposes a new approach to solve electromagnetic design problems that combines the DE algorithm with a generator of chaos sequences. This approach is tested on the design of a loudspeaker model with 17 degrees of freedom, for showing its applicability to electromagnetic problems. The results show that the DE algorithm with chaotic sequences presents better, or at least similar, results when compared to the standard DE algorithm and other evolutionary algorithms available in the literature.

New Optimization Algorithms for Structural Reliability Analysis

Solution of structural reliability problems by the First Order method require optimization algorithms to find the smallest distance between a limit state function and the origin of standard Gaussian space. The Hassofer-Lind-Rackwitz-Fiessler (HLRF) algorithm, developed specifically for this purpose, has been shown to be efficient but not robust, as it fails to converge for a significant number of problems. On the other hand, recent developments in general (augmented Lagrangian) optimization techniques have not been tested in aplication to structural reliability problems. In the present article, three new optimization algorithms for structural reliability analysis are presented. One algorithm is based on the HLRF, but uses a new differentiable merit function with Wolfe conditions to select step length in linear search. It is shown in the article that, under certain assumptions, the proposed algorithm generates a sequence that converges to the local minimizer of the problem. Two new augmented Lagrangian methods are also presented, which use quadratic penalties to solve nonlinear problems with equality constraints. Performance and robustness of the new algorithms is compared to the classic augmented Lagrangian method, to HLRF and to the improved HLRF (iHLRF) algorithms, in the solution of 25 benchmark problems from the literature. The new proposed HLRF algorithm is shown to be more robust than HLRF or iHLRF, and as efficient as the iHLRF algorithm. The two augmented Lagrangian methods proposed herein are shown to be more robust and more efficient than the classical augmented Lagrangian method.

Design of laminated piezocomposite shell transducers with arbitrary fiber orientation using topology optimization approach

Sensor and actuator based on laminated piezocomposite shells have shown increasing demand in the field of smart structures. The distribution of piezoelectric material within material layers affects the performance of these structures; therefore, its amount, shape, size, placement, and polarization should be simultaneously considered in an optimization problem. In addition, previous works suggest the concept of laminated piezocomposite structure that includes fiber-reinforced composite layer can increase the performance of these piezoelectric transducers; however, the design optimization of these devices has not been fully explored yet. Thus, this work aims the development of a methodology using topology optimization techniques for static design of laminated piezocomposite shell structures by considering the optimization of piezoelectric material and polarization distributions together with the optimization of the fiber angle of the composite orthotropic layers, which is free to assume different values along the same composite layer. The finite element model is based on the laminated piezoelectric shell theory, using the degenerate three-dimensional solid approach and first-order shell theory kinematics that accounts for the transverse shear deformation and rotary inertia effects. The topology optimization formulation is implemented by combining the piezoelectric material with penalization and polarization model and the discrete material optimization, where the design variables describe the amount of piezoelectric material and polarization sign at each finite element, with the fiber angles, respectively. Three different objective functions are formulated for the design of actuators, sensors, and energy harvesters. Results of laminated piezocomposite shell transducers are presented to illustrate the method. Copyright (C) 2012 John Wiley & Sons, Ltd.

Models and Algorihtm for the Optimization of Real-World Routing and Logistics Problems

Logistics involves planning, managing, and organizing the flows of goods from the point of origin to the point of destination in order to meet some requirements. Logistics and transportation aspects are very important and represent a relevant costs for producing and shipping companies, but also for public administration and private citizens. The optimization of resources and the improvement in the organization of operations is crucial for all branches of logistics, from the operation management to the transportation. As we will have the chance to see in this work, optimization techniques, models, and algorithms represent important methods to solve the always new and more complex problems arising in different segments of logistics. Many operation management and transportation problems are related to the optimization class of problems called Vehicle Routing Problems (VRPs). In this work, we consider several real-world deterministic and stochastic problems that are included in the wide class of the VRPs, and we solve them by means of exact and heuristic methods. We treat three classes of real-world routing and logistics problems. We deal with one of the most important tactical problems that arises in the managing of the bike sharing systems, that is the Bike sharing Rebalancing Problem (BRP). We propose models and algorithms for real-world earthwork optimization problems. We describe the 3DP process and we highlight several optimization issues in 3DP. Among those, we define the problem related to the tool path definition in the 3DP process, the 3D Routing Problem (3DRP), which is a generalization of the arc routing problem. We present an ILP model and several heuristic algorithms to solve the 3DRP.

Towards description and optimization of abstract machines in an extension of prolog

Competitive abstract machines for Prolog are usually large, intricate, and incorpórate sophisticated optimizations. This makes them difñcult to code, optimize, and, especially, maintain and extend. This is partly due to the fact that efñciency considerations make it necessary to use low-level languages in their implementation. Writing the abstract machine (and ancillary code) in a higher-level language can help harness this inherent complexity. In this paper we show how the semantics of basic components of an efficient virtual machine for Prolog can be described using (a variant of) Prolog which retains much of its semantics. These descriptions are then compiled to C and assembled to build a complete bytecode emulator. Thanks to the high level of the language used and its closeness to Prolog the abstract machine descriptions can be manipulated using standard Prolog compilation and optimization techniques with relative ease. We also show how, by applying program transformations selectively, we obtain abstract machine implementations whose performance can match and even exceed that of highly-tuned, hand-crafted emulators.

Methods and Techniques for the Generation and Eficient Exploitation of RDB2RDF Mappings

RDB to RDF Mapping Language (R2RML) es una recomendación del W3C que permite especificar reglas para transformar bases de datos relacionales a RDF. Estos datos en RDF se pueden materializar y almacenar en un sistema gestor de tripletas RDF (normalmente conocidos con el nombre triple store), en el cual se pueden evaluar consultas SPARQL. Sin embargo, hay casos en los cuales la materialización no es adecuada o posible, por ejemplo, cuando la base de datos se actualiza frecuentemente. En estos casos, lo mejor es considerar los datos en RDF como datos virtuales, de tal manera que las consultas SPARQL anteriormente mencionadas se traduzcan a consultas SQL que se pueden evaluar sobre los sistemas gestores de bases de datos relacionales (SGBD) originales. Para esta traducción se tienen en cuenta los mapeos R2RML. La primera parte de esta tesis se centra en la traducción de consultas. Se propone una formalización de la traducción de SPARQL a SQL utilizando mapeos R2RML. Además se proponen varias técnicas de optimización para generar consultas SQL que son más eficientes cuando son evaluadas en sistemas gestores de bases de datos relacionales. Este enfoque se evalúa mediante un benchmark sintético y varios casos reales. Otra recomendación relacionada con R2RML es la conocida como Direct Mapping (DM), que establece reglas fijas para la transformación de datos relacionales a RDF. A pesar de que ambas recomendaciones se publicaron al mismo tiempo, en septiembre de 2012, todavía no se ha realizado un estudio formal sobre la relación entre ellas. Por tanto, la segunda parte de esta tesis se centra en el estudio de la relación entre R2RML y DM. Se divide este estudio en dos partes: de R2RML a DM, y de DM a R2RML. En el primer caso, se estudia un fragmento de R2RML que tiene la misma expresividad que DM. En el segundo caso, se representan las reglas de DM como mapeos R2RML, y también se añade la semántica implícita (relaciones de subclase, 1-N y M-N) que se puede encontrar codificada en la base de datos. Esta tesis muestra que es posible usar R2RML en casos reales, sin necesidad de realizar materializaciones de los datos, puesto que las consultas SQL generadas son suficientemente eficientes cuando son evaluadas en el sistema gestor de base de datos relacional. Asimismo, esta tesis profundiza en el entendimiento de la relación existente entre las dos recomendaciones del W3C, algo que no había sido estudiado con anterioridad. ABSTRACT. RDB to RDF Mapping Language (R2RML) is a W3C recommendation that allows specifying rules for transforming relational databases into RDF. This RDF data can be materialized and stored in a triple store, so that SPARQL queries can be evaluated by the triple store. However, there are several cases where materialization is not adequate or possible, for example, if the underlying relational database is updated frequently. In those cases, RDF data is better kept virtual, and hence SPARQL queries over it have to be translated into SQL queries to the underlying relational database system considering that the translation process has to take into account the specified R2RML mappings. The first part of this thesis focuses on query translation. We discuss the formalization of the translation from SPARQL to SQL queries that takes into account R2RML mappings. Furthermore, we propose several optimization techniques so that the translation procedure generates SQL queries that can be evaluated more efficiently over the underlying databases. We evaluate our approach using a synthetic benchmark and several real cases, and show positive results that we obtained. Direct Mapping (DM) is another W3C recommendation for the generation of RDF data from relational databases. While R2RML allows users to specify their own transformation rules, DM establishes fixed transformation rules. Although both recommendations were published at the same time, September 2012, there has not been any study regarding the relationship between them. The second part of this thesis focuses on the study of the relationship between R2RML and DM. We divide this study into two directions: from R2RML to DM, and from DM to R2RML. From R2RML to DM, we study a fragment of R2RML having the same expressive power than DM. From DM to R2RML, we represent DM transformation rules as R2RML mappings, and also add the implicit semantics encoded in databases, such as subclass, 1-N and N-N relationships. This thesis shows that by formalizing and optimizing R2RML-based SPARQL to SQL query translation, it is possible to use R2RML engines in real cases as the resulting SQL is efficient enough to be evaluated by the underlying relational databases. In addition to that, this thesis facilitates the understanding of bidirectional relationship between the two W3C recommendations, something that had not been studied before.