Dissertationen zum Thema „Logic optimizations“

Tato práce se zabývá metodami logické syntézy a optimalizací pro polymorfní obvody. V práci jsou jak diskutovány existující metody pro konvenční obvody, tak i představeny nové metody, aplikovatelné na polymorfní elektroniku. Hlavním přínosem práce je představení nových metod optimalizace a logické syntézy pro polymorfní obvody. Přesto, že v minulých letech byly představeny metody pro návrh polymorfních obvodů, jsou tyto metody založené na evolučních technikách nebo nejsou dobře škálovatelné. Z toho vyplývá, že stále neexistuje stabilní metodika pro návrh složitějších polymorfních obvodů. Tato práce představuje zejména reprezentaci polymorgních obvodů a metodiku pro jejich návrh založenou na And-Inverter grafech. Na polymorfní obvody reprezentované pomocí AIG je možné aplikovat známé techniky jako například přepisování [rewriting]. Nasazením techniky přepisování na polymorfní AIG získáme obvod, obsahující polymorfní prvky uvnitř obvodu, a je možné dosáhnout značných úspor prostředků, které mohou být sdíleny mezi dvěma funkcemi současně. Ověření návrhové metodiky pro polymorfní obvody bylo provedeno nad sadou veřejně dostupných obvodů, čímž je demonstrována efektivita metodiky.

Gate level simulation is a necessary step to verify the correctness of a circuitdesign before fabrication. It is a very time-consuming application, especially in lightof current circuit sizes. Since circuits are continually growing in size and complexity,there is a need for more efficient simulation techniques to keep the circuit verificationtime acceptably small. The use of parallel or distributed simulation is such a technique.When executed on a network of workstations, distributed simulation is alsoa very cost-effective technique. This research focuses on optimization techniques forTime Warp based gate-level logic simulations. The techniques which are described inthis thesis are oriented towards distributed platforms. The first major contributionof this thesis was the creation of an object oriented distributed simulator, XTW. Ituses an optimistic synchronization algorithm and incorporates a number of knownoptimization techniques targeting different aspects of distributed logic simulation.XEQ, an O(1) event scheduling algorithm for this simulator was developed for usein XTW. XEQ enabled us to execute gate level simulations up to 9.4 times fasterthan the same simulator using a skip-list (O(lg n)) event queue. rb-messagea mechanism which reduces the cost of rollback in Time Warp was also developedfor use in XTW. Our experiments revealed that the rb-message mechanism reducedthe number of anti-messages sent in a Time Warp based logic simulation by 76%on average. Moreover, based on the observations that (1)not all circuits should besimulated in parallel and (2) different circuits achieve their best parallel simulationperformance with a different number of compute nodes, an algorithm that uses theK-NN machine learning algorithm was devised to determine the most effective softwareand hardware combination for a logic simulation. After an extensive trainingregime, it was shown to make a correct prediction 99% of the time on whether touse a parallel or sequential simulator. The predicted number of nodes to use on aparallel platform was shown to produce an average execution time which was notmore than 12% of the smallest execution time. The configuration which resulted inthe minimal execution time was picked 61% of the time. A final contribution of thisthesis is an effort to link together commercial single processor simulators making useof Verilog PLI.
La simulation "gate-level" est une tape ncessaire pour vrifier la conformit dela conception d'un circuit avant sa fabrication. C'est un programme qui prendbeaucoup de temps, compte tenu particulirement de la taille actuelle des circuits.Ceux-ci ne cessant de se dvelopper en taille et en complexit, il y a un rel besoin detechniques de simulation plus efficaces afin de maintenir la dure de vrification ducircuit raisonnablement courte. Une de ces techniques consiste utiliser la simulationparallle ou distribue. Quand excute sur un rseau de postes de travail, la simulationdistribue se rvle galement tre une technique trs rentable. Cette recherche se concentresur l'optimisation des techniques de simulations "gate-level" logiques bases surTime Warp. Les techniques qui sont dcrites dans cet expos sont orientes vers lesplateformes distribues. La premire contribution majeure de cet expos a t la crationd'un simulateur distribu orient sur l'objet, XTW. Il utilise un algorithme de synchronisationoptimiste et incorpore un certain nombre de techniques d'optimisationconnues visant diffrents aspects de la simulation distribue logique. XEQ, un algorithmeprogrammateur d'vnements O(1) pour ce simulateur a t dvelopp pour treutilis dans XTW. XEQ nous permet d'excuter des simulations "gate-level" jusqu'9,4 fois plus rapides qu'avec le mme simulateur utilisant une suite d'vnement en"skip-list" (O(lg n)). "rb-message" – un mcanisme qui diminue le co?t de rductiondans Time Warp a galement t mis au point pour tre utilis dans XTW. Nos essaisont rvl que le mcanisme de "rb-message" permettait de diminuer le nombre des antimessagesenvoys au cours d'une simulation logique base sur Time Warp de 76 % enmoyenne. Il a t en outre con?u, en se basant sur les observations que (1) certainscircuits ne devraient pas tre simuls en parallle et (2) que diffrents circuits atteignentleur meilleure performance de simulation parallle avec un nombre diffrent de noeudsde calculs, un algorithme utilisant l'algorithme d'apprentissage de la machine K-NNafin de dterminer quelle tait l'association de logiciel et de matriel la plus efficacedans le cadre d'une simulation logique. l'issue d'un entra?nement approfondi, ilest apparu qu'il pouvait faire un pronostic juste 99 % tablissant quand utiliser unsimulateur parallle ou squentiel. Le nombre annonc de noeuds utiliser sur une plateformeparallle s'est avr permettre une dure d'excution moyenne gale 12 % de la pluscourte dure d'excution. La configuration ayant abouti la dure d'excution minimalea t reprise dans 61 % des cas. Dernire contribution apporte par cet expos, relier lessimulateurs commerciaux processeur unique utilisant Verilog PLI.

Fuzzy logic systems are widely used for control, system identification, and pattern recognition problems. In order to maximize their performance, it is often necessary to undertake a design optimization process in which the adjustable parameters defining a particular fuzzy system are tuned to maximize a given performance criterion. Some data to approximate are commonly available and yield what is called the supervised learning problem. In this problem we typically wish to minimize the sum of the squares of errors in approximating the data. We first introduce fuzzy logic systems and the supervised learning problem that, in effect, is a nonlinear optimization problem that at times can be non-differentiable. We review the existing approaches and discuss their weaknesses and the issues involved. We then focus on one of these problems, i.e., non-differentiability of the objective function, and show how current approaches that do not account for non-differentiability can diverge. Moreover, we also show that non-differentiability may also have an adverse practical impact on algorithmic performances. We reformulate both the supervised learning problem and piecewise linear membership functions in order to obtain a polynomial or factorable optimization problem. We propose the application of a global nonconvex optimization approach, namely, a reformulation and linearization technique. The expanded problem dimensionality does not make this approach feasible at this time, even though this reformulation along with the proposed technique still bears a theoretical interest. Moreover, some future research directions are identified. We propose a novel approach to step-size selection in batch training. This approach uses a limited memory quadratic fit on past convergence data. Thus, it is similar to response surface methodologies, but it differs from them in the type of data that are used to fit the model, that is, already available data from the history of the algorithm are used instead of data obtained according to an experimental design. The step-size along the update direction (e.g., negative gradient or deflected negative gradient) is chosen according to a criterion of minimum distance from the vertex of the quadratic model. This approach rescales the complexity in the step-size selection from the order of the (large) number of training data, as in the case of exact line searches, to the order of the number of parameters (generally lower than the number of training data). The quadratic fit approach and a reduced variant are tested on some function approximation examples yielding distributions of the final mean square errors that are improved (i.e., skewed toward lower errors) with respect to the ones in the commonly used pattern-by-pattern approach. Moreover, the quadratic fit is also competitive and sometimes better than the batch training with optimal step-sizes, thus showing an improved performance of this approach. The quadratic fit approach is also tested in conjunction with gradient deflection strategies and memoryless variable metric methods, showing errors smaller by 1 to 7 orders of magnitude. Moreover, the convergence speed by using either the negative gradient direction or a deflected direction is higher than that of the pattern-by-pattern approach, although the computational cost of the algorithm per iteration is moderately higher than the one of the pattern-by-pattern method. Finally, some directions for future research are identified.
Ph. D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Mechanical Engineering, 2005.
Includes bibliographical references (p. 109-110).
We present an input-output model that learns to emulate a complex building simulation of high dimensionality. Many multi-dimensional systems are dominated by the behavior of a small number of inputs over a limited range of input variation. Some also exhibit a tendency to respond relatively strongly to certain inputs over small ranges, and to other inputs over very large ranges of input variation. A branching linear discriminant can be used to isolate regions of local linearity in the input space, while also capturing the effects of scale. The quality of the classification may be improved by using a fuzzy preference relation to classify input configurations that are not well handled by the linear discriminant.
by Matthew A. Lehar.
Ph.D.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1994.
Includes bibliographical references (leaves 69-71).
by Mazhar Murtaza Alidina.
M.S.

With the increased complexity of Very Large Scaled Integrated (VLSI) circuits, multilevel logic synthesis plays an even more important role due to its flexibility and compactness. The history of symbolic logic and some typical techniques for multilevel logic synthesis are reviewed. These methods include algorithmic approach; Rule-Based approach; Binary Decision Diagram (BDD) approach; Field Programmable Gate Array(FPGA) approach and several perturbation applications. One new kind of don't cares (DCs), called functional DCs has been proposed for multilevel logic synthesis. The conventional two-level cubes are generalized to multilevel cubes. Then functional DCs are generated based on the properties of containment. The concept of containment is more general than unateness which leads to the generation of new DCs. A separate C program has been developed to utilize the functional DCs generated as a Boolean function is decomposed for both single output and multiple output functions. The program can produce better results than script.rugged of SIS, developed by UC Berkeley, both in area and speed in less CPU time for a number of testcases from MCNC and IWLS'93 benchmarks. In certain applications ANDjXOR (Reed-Muller) logic has shown some attractive advantages over the standard Boolean logic based on AND JOR operations. A bidirectional conversion algorithm between these two paradigms is presented based on the concept of polarity for sum-of-products (SOP) Boolean functions, multiple segment and multiple pointer facilities. Experimental results show that the algorithm is much faster than the previously published programs for any fixed polarity. Based on this algorithm, a new technique called redundancy-removal is applied to generalize the idea to very large multiple output Boolean functions. Results for benchmarks with up to 199 inputs and 99 outputs are presented. Applying the preceding conversion program, any Boolean functions can be expressed by fixed polarity Reed-Muller forms. There are 2n polarities for an n-variable function and the number of product terms depends on these polarities. The problem of exact polarity minimization is computationally extensive and current programs are only suitable when n :::; 15. Based on the comparison of the concepts of polarity in the standard Boolean logic and Reed-Muller logic, a fast algorithm is developed and implemented in C language which can find the best polarity for multiple output functions. Benchmark examples of up to 25 inputs and 29 outputs run on a personal computer are given. After the best polarity for a Boolean function is calculated, this function can be further simplified using mixed polarity methods by combining the adjacent product terms. Hence, an efficient program is developed based on decomposition strategy to implement mixed polarity minimization for both single output and very large multiple output Boolean functions. Experimental results show that the numbers of product terms are much less than the results produced by ESPRESSO for some categories of functions.

Esta tesis propone una metodología de análisis dinámico que mejora el diagnóstico de programas erróneos escritos en el lenguaje Maude. La idea clave es combinar técnicas de verificación de aserciones en tiempo de ejecución con la fragmentación dinámica de trazas de ejecución para detectar automáticamente errores en tiempo de ejecución, al tiempo que se reduce el tamaño y la complejidad de las trazas a analizar. En el caso de violarse una aserción, se infiere automáticamente el criterio de fragmentación, lo que facilita al usuario identificar rápidamente la fuente del error. En primer lugar, la tesis formaliza una técnica destinada a detectar automáticamente eventuales desviaciones del comportamiento deseado del programa (síntomas de error). Esta técnica soporta dos tipos de aserciones definidas por el usuario: aserciones funcionales (que restringen llamadas a funciones deterministas) y aserciones de sistema (que especifican los invariantes de estado del sistema). La técnica de verificación dinámica propuesta es demostrablemente correcta en el sentido de que todos los errores señalados definitivamente delatan la violación de las aserciones. Tras eventuales violaciones de aserciones, se generan automáticamente trazas fragmentadas (es decir, trazas simplificadas pero igualmente precisas) que ayudan a identificar la causa del error. Además, la técnica también sugiere una posible reparación para las reglas implicadas en la generación de los estados erróneos. La metodología propuesta se basa en (i) una notación lógica para especificar las aserciones que se imponen a la ejecución; (ii) una técnica de verificación aplicable en tiempo de ejecución que comprueba dinámicamente las aserciones; y (iii) un mecanismo basado en la generalización (ecuacional) menos general que automáticamente obtiene criterios precisos para fragmentar trazas de ejecución a partir de aserciones falsificadas. Por último, se presenta una implementación de la técnica propuesta en la herramienta de análisis dinámico basado en aserciones ABETS, que muestra cómo es posible combinar el trazado de las propiedades asertadas del programa para obtener un algoritmo preciso de análisis de trazas que resulta útil para el diagnóstico y la depuración de programas.
This thesis proposes a dynamic analysis methodology for improving the diagnosis of erroneous Maude programs. The key idea is to combine runtime assertion checking and dynamic trace slicing for automatically catching errors at runtime while reducing the size and complexity of the erroneous traces to be analyzed (i.e., those leading to states that fail to satisfy the assertions). In the event of an assertion violation, the slicing criterion is automatically inferred, which facilitates the user to rapidly pinpoint the source of the error. First, a technique is formalized that aims at automatically detecting anomalous deviations of the intended program behavior (error symptoms) by using assertions that are checked at runtime. This technique supports two types of user-defined assertions: functional assertions (which constrain deterministic function calls) and system assertions (which specify system state invariants). The proposed dynamic checking is provably sound in the sense that all errors flagged definitely signal a violation of the specifications. Then, upon eventual assertion violations, accurate trace slices (i.e., simplified yet precise execution traces) are generated automatically, which help identify the cause of the error. Moreover, the technique also suggests a possible repair for the rules involved in the generation of the erroneous states. The proposed methodology is based on (i) a logical notation for specifying assertions that are imposed on execution runs; (ii) a runtime checking technique that dynamically tests the assertions; and (iii) a mechanism based on (equational) least general generalization that automatically derives accurate criteria for slicing from falsified assertions. Finally, an implementation of the proposed technique is presented in the assertion-based, dynamic analyzer ABETS, which shows how the forward and backward tracking of asserted program properties leads to a thorough trace analysis algorithm that can be used for program diagnosis and debugging.
Esta tesi proposa una metodologia d'anàlisi dinàmica que millora el diagnòstic de programes erronis escrits en el llenguatge Maude. La idea clau és combinar tècniques de verificació d'assercions en temps d'execució amb la fragmentació dinàmica de traces d'execució per a detectar automàticament errors en temps d'execució, alhora que es reduïx la grandària i la complexitat de les traces a analitzar. En el cas de violar-se una asserció, s'inferix automàticament el criteri de fragmentació, la qual cosa facilita a l'usuari identificar ràpidament la font de l'error. En primer lloc, la tesi formalitza una tècnica destinada a detectar automàticament eventuals desviacions del comportament desitjat del programa (símptomes d'error). Esta tècnica suporta dos tipus d'assercions definides per l'usuari: assercions funcionals (que restringixen crides a funcions deterministes) i assercions de sistema (que especifiquen els invariants d'estat del sistema). La tècnica de verificació dinàmica proposta és demostrablement correcta en el sentit que tots els errors assenyalats definitivament delaten la violació de les assercions. Davant eventuals violacions d'assercions, es generen automàticament traces fragmentades (és a dir, traces simplificades però igualment precises) que ajuden a identificar la causa de l'error. A més, la tècnica també suggerix una possible reparació de les regles implicades en la generació dels estats erronis. La metodologia proposada es basa en (i) una notació lògica per a especificar les assercions que s'imposen a l'execució; (ii) una tècnica de verificació aplicable en temps d'execució que comprova dinàmicament les assercions; i (iii) un mecanisme basat en la generalització (ecuacional) menys general que automàticament obté criteris precisos per a fragmentar traces d'execució a partir d'assercions falsificades. Finalment, es presenta una implementació de la tècnica proposta en la ferramenta d'anàlisi dinàmica basat en assercions ABETS, que mostra com és possible combinar el traçat cap avant i cap arrere de les propietats assertades del programa per a obtindre un algoritme precís d'anàlisi de traces que resulta útil per al diagnòstic i la depuració de programes.
Sapiña Sanchis, J. (2017). Rewriting Logic Techniques for Program Analysis and Optimization [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/94044
TESIS

Inspired by the idea of vision zero, there is a lot of work that needs to be done in the field of advance driver assistance systems to develop more safer systems. Driver intention detection with a prediction of upcoming behavior of the driver is one possible solution to reduce the fatalities in road traffic. Driver intention detection provides an early warning of the driver's behavior to an Advanced Driver Assistance Systems (ADAS) and at the same time reduces the risk of non-essential warnings. This will significantly reduce the problem of warning dilemma and the system will become more safer. A driving maneuver prediction can be regarded as an implementation of driver's behavior. So the aim of this thesis is to determine the driver's intention by early prediction of a driving maneuver using Controller Area Network (CAN) bus data. The focus of this thesis is to optimize and further develop an algorithm for driver intention detection with fuzzy logic and edit distance method. At first the basics concerning driver's intention detection are described as there exists different ways to determine it. This work basically uses CAN bus data to determine a driver's intention. The algorithm overview with the design parameters are described next to have an idea about the functioning of the algorithm. Then different implementation tasks are explained for optimization and further development of the algorithm. The main aim to execute these implementation tasks is to improve the overall performance of the algorithm concerning True Positive Rate (TPR), False Positive Rate (FPR) and earliness values. At the end, the results are validated to check the algorithm performance with different possibilities and a test drive is performed to evaluate the real time capability of the algorithm. Lastly the use of driver intention detection algorithm for an ADAS to make it more safer is described in details. The early warning information can be feed to an ADAS, for example, an automatic collision avoidance or a lane change assistance ADAS to further improve safety for these systems.

Genetic algorithms are commonly used to solve combinatorial optimizationproblems. The implementation evolves using genetic operators (crossover, mutation,selection, etc.). Anyway, genetic algorithms like some other methods have parameters(population size, probabilities of crossover and mutation) which need to be tune orchosen.In this paper, our project is based on an existing hybrid genetic algorithmworking on the multiprocessor scheduling problem. We propose a hybrid Fuzzy-Genetic Algorithm (FLGA) approach to solve the multiprocessor scheduling problem.The algorithm consists in adding a fuzzy logic controller to control and tunedynamically different parameters (probabilities of crossover and mutation), in anattempt to improve the algorithm performance. For this purpose, we will design afuzzy logic controller based on fuzzy rules to control the probabilities of crossoverand mutation. Compared with the Standard Genetic Algorithm (SGA), the resultsclearly demonstrate that the FLGA method performs significantly better.

Today it is possible to integrate more than one billion transistors onto a single chip. This has enabled implementation of complex functionality in hand held gadgets, but handling such complexity is far from trivial. The challenges of handling this complexity are mostly related to the design and testing of the digital components of these chips. A number of well-researched disciplines must be employed in the efficient design of large and complex chips. These include utilization of several abstraction levels, design of appropriate architectures, several different classes of optimization methods, and development of testing techniques. This thesis contributes mainly to the areas of design optimization and testing methods. In the area of testing this thesis contributes methods for testing of on-chip links connecting different clock domains. This includes testing for defects that introduce unacceptable delay, lead to excessive crosstalk and cause glitches, which can produce errors. We show how pure digital components can be used to detect such defects and how the tests can be scheduled efficiently. To manage increasing test complexity, another contribution proposes to raise theabstraction level of fault models from logic level to system level. A set of system level faultmodels for a NoC-switch is proposed and evaluated to demonstrate their potential. In the area of design optimization, this thesis focuses primarily on logic optimization. Two contributions for Boolean decomposition are presented. The first one is a fast heuristic algorithm that finds non-disjoint decompositions for Boolean functions. This algorithm operates on a Binary Decision Diagram. The other contribution is a fast algorithm for detecting whether a function is likely to benefit from optimization for architectures with a gate depth of three with an XOR-gate as the third gate.

An Optimization Approach to Employee Scheduling Using Fuzzy Logic William G. Spence Selection of sales employees is critical because the sales employees represent the company’s image, competitive advantage, technology, and values. In many service systems the majority of consumer contact is with the sales department. Since there are different types of customers, scheduling quality salespersons who can adequately help consumers may affect revenue. This thesis proposes a new methodology for the scheduling of employees in a service system. The methodology uses Fuzzy Logic to calculate possible sales and Linear Programming to create an optimal schedule. This approach enables the rating of sales employees with respect to three customer’s types (Lookie Lou, Price Shopper and Buyer). The salesperson rating, along with customer arrival distribution is then used to optimize sale person scheduling, with the objective of revenue maximization. The uniqueness of this thesis lies in the combination of Fuzzy Logic and Linear Programming. The combination of these two disciplines provides an adaptive tool that can be used to optimize employee scheduling based on personality traits.

Description Logics (DLs) form a family of languages which correspond to decidable fragments of First-Order Logic (FOL). They have been overwhelmingly successful for constructing ontologies - conceptual structures describing domain knowledge. Ontologies proved to be valuable in a range of areas, most notably, bioinformatics, chemistry, Health Care and Life Sciences, and the Semantic Web.One limitation of DLs, as fragments of FOL, is their restricted ability to cope with various forms of uncertainty. For example, medical knowledge often includes statistical relationships, e.g., findings or results of clinical trials. Currently it is maintained separately, e.g., in Bayesian networks or statistical models. This often hinders knowledge integration and reuse, leads to duplication and, consequently, inconsistencies.One answer to this issue is probabilistic logics which allow for smooth integration of classical, i.e., expressible in standard FOL or its sub-languages, and uncertain knowledge. However, probabilistic logics have long been considered impractical because of discouraging computational properties. Those are mostly due to the lack of simplifying assumptions, e.g., independence assumptions which are central to Bayesian networks.In this thesis we demonstrate that deductive reasoning in a particular probabilistic DL, called P-SROIQ, can be computationally practical. We present a range of novel algorithms, in particular, the probabilistic satisfiability procedure (PSAT) which is, to our knowledge, the first scalable PSAT algorithm for a non-propositional probabilistic logic. We perform an extensive performance and scalability evaluation on different synthetic and natural data sets to justify practicality.In addition, we study theoretical properties of P-SROIQ by formally translating it into a fragment of first-order logic of probability. That allows us to gain a better insight into certain important limitations of P-SROIQ. Finally, we investigate its applicability from the practical perspective, for instance, use it to extract all inconsistencies from a real rule-based medical expert system.We believe the thesis will be of interest to developers of probabilistic reasoners. Some of the algorithms, e.g., PSAT, could also be valuable to the Operations Research community since they are heavily based on mathematical programming. Finally, the theoretical analysis could be helpful for designers of future probabilistic logics.

The workshops on (constraint) logic programming (WLP) are the annual meeting of the Society of Logic Programming (GLP e.V.) and bring together researchers interested in logic programming, constraint programming, and related areas like databases, artificial intelligence and operations research. In this decade, previous workshops took place in Dresden (2008), Würzburg (2007), Vienna (2006), Ulm (2005), Potsdam (2004), Dresden (2002), Kiel (2001), and Würzburg (2000). Contributions to workshops deal with all theoretical, experimental, and application aspects of constraint programming (CP) and logic programming (LP), including foundations of constraint/ logic programming. Some of the special topics are constraint solving and optimization, extensions of functional logic programming, deductive databases, data mining, nonmonotonic reasoning,
interaction of CP/LP with other formalisms like agents, XML, JAVA, program analysis, program transformation, program verification, meta programming, parallelism and concurrency, answer set programming, implementation and software techniques (e.g., types, modularity, design patterns), applications (e.g., in production, environment, education, internet), constraint/logic programming for semantic web systems and applications, reasoning on the semantic web, data modelling for the web, semistructured data, and web query languages.

Specialty chemicals industry relies on batch manufacturing, since it requires the frequent adaptation of production systems to market fluctuations. To be first in the market, batch industry requires decision-support systems for the rapid development and implementation of chemical processes. Moreover, the processes should be competitive to ensure their long-term viability. General-purpose and flexible plants and the consideration of physicochemical insights to define an efficient operation are also cornerstones for the success of specialty chemical industries. Precisely, this thesis tackles the systematic development of batch processes that are efficient, economically competitive, and environmentally friendly, to assist their agile introduction into production systems in grassroots and retrofit scenarios. Synthesis of conceptual processing schemes and plant allocation subproblems are solved simultaneously, taking into account the plant design. With this purpose, an optimization-based approach is proposed, where all structural alternatives are represented in a State-Equipment Network (SEN) superstructure, following formulated into a Mixed-Logic Dynamic Optimization (MLDO) problem which is later solved to minimize an objective function. Essentially, the strength of the proposed methodology lies in the modeling strategy which combines the different kinds of decisions of the integrated problem in a unique optimization model. Accordingly, it considers: (i) synthesis and allocation alternatives combination, (ii) dynamic process performance models and dynamic control variable profiles, (iii) discrete events associated to transitions of batch phases and operations, (iv) quantitative and qualitative information, (v) material transference synchronization to ensure batch integrity between unit procedures, and (vi) batch and semicontinuous processing elements. Different strategies can be used to solve the resulting MLDO problem. A deterministic direct-simultaneous approach is first proposed. The mixed-logic problem is reformulated into a mixed-integer one, which is fully-discretized to provide a Mixed-Integer Non-Linear Programming (MINLP) that is optimized using conventional solvers. Then, a Differential Genetic Algorithm (DGA) and a hybrid approach are presented. The purpose of these evolutionary strategies is to pose solution alternatives that keep solution goodness while seek for the improvement of computational efficiency to handle industrial-size problems. The optimization-based approach is applied in retrofit scenarios to solve the simultaneous process synthesis and plant allocation, taking into account the physical restrictions of existing plant elements. The production of specialty chemicals based on a competitive reactions system in an existing reactor network is first defined through process development and improvement according to different economic scenarios, decision criteria, and plant modifications. Additionally, a photo-Fenton process is optimized to eliminate an emergent wastewater pollutant in a given pilot plant, pursuing the minimization of processing time and cost. Batch process development in grassroots scenarios is also proven to be a problem of utmost importance to deal with uncertainty in future markets. Seeking for plant flexibility in several demand scenarios, the expected profit is maximized through a two-stage stochastic formulation that includes simultaneous plant design, process synthesis, and plant allocation decisions. A heuristic solution algorithm is used to handle the problem complexity. A grassroots plant design is defined to implement the previous competitive reaction system, where decisions like the feed-forward trajectories or operating modes allow the adaptation of master recipes to different demands. Finally, an acrylic fiber production example is presented to illustrate process development decisions like the selection of tasks, technological alternatives, chemicals, and solvent reuse.
La indústria de productes químics especials es basa en la fabricació discontinua, ja que permet adaptar de forma freqüent els sistemes de producció en funció de les fluctuacions de mercat. Per ser líder al sector, són necessàries eines de suport a la decisió que ajudin a l’àgil desenvolupament i implementació de nous processos. A més, aquests han de ser competitius per garantir la seva viabilitat a llarg termini. Altres peces clau per una operació eficient són l’ús de plantes flexibles així com l’estudi dels fenòmens fisicoquímics. Aquesta tesis aborda justament el desenvolupament sistemàtic de processos químics discontinus que siguin eficients, econòmicament competitius i ecològics, per contribuir a la seva ràpida introducció en els sistemes de producció, tant en escenaris de plantes existents com des de les bases. En concret, es planteja la resolució simultània de la síntesi conceptual d’esquemes de procés i l’assignació d’equips, tenint en compte el disseny de la planta. Amb aquest objectiu, es proposa una metodologia de solució basada en optimització, on les alternatives estructurals es representen en una Xarxa d’Estats i Equips (SEN per les sigles en anglès) que es formula mitjançant un problema d’Optimització Dinàmica Mixta-Lògica (MLDO per les sigles en anglès) que es resol minimitzant una funció objectiu. La solidesa de la metodologia proposada rau en la estratègia de modelat del problema MLDO, que integra els diferents tipus de decisions en un sol model d’optimització. En concret, es consideren: (i) la combinació d’alternatives de síntesi i assignació d’equips, (ii) models de procés i trajectòries de control dinàmics, (iii) esdeveniments discrets associats al canvi de fase i operació, (iv) informació quantitativa i qualitativa, (v) sincronització de transferències de material en tasques consecutives, i (vi) elements de processat discontinus i semi-continus. Existeixen diverses estratègies per resoldre el problema MLDO resultant. En aquesta tesi es proposa en primer lloc un mètode determinístic directe-simultani, on el model mixt-lògic es transforma en un mixt-enter. Aquest es discretitza al seu torn de forma completa per obtenir un problema de Programació No-Lineal Mixta-Entera (MINLP per les sigles en anglès) el qual es pot resoldre utilitzant algoritmes d’optimització convencionals. A més, es presenten un Algoritme Genètic Diferencial (DGA per les sigles en anglès) i un mètode híbrid. Totes dues estratègies esdevenen alternatives de cerca amb l’objectiu de mantenir la bondat de la solució i millorar l’eficàcia de computació per tractar problemes de dimensió industrial. La metodologia de solució proposada s’aplica al desenvolupament de processos discontinus en escenaris de plantes existents, tenint en compte les restriccions físiques dels equips. Un primer exemple aborda la manufactura de productes químics basada en un sistema de reaccions competitives. Concretament, es desenvolupa i millora el procés de producció implementat en una xarxa de reactors considerant diferents escenaris econòmics, criteris de decisió, i modificacions de planta. En un segon exemple, s’optimitza el procés foto-Fenton per ser executat en una planta pilot per eliminar contaminants emergents. Buscant integrar el desenvolupament de procés i el disseny de plantes flexibles en escenaris de base, es presenta una formulació estocàstica en dues etapes per a optimitzar el benefici esperat d’acord a diversos escenaris de demanda. Per gestionar la complexitat d’aquest problema es proposa la utilització d’una heurística. Com a exemple, es planteja el disseny d’una planta de base on implementar l’anterior sistema de reaccions competitives. Decisions com les trajectòries dinàmiques de control o la configuració d’equips permeten adaptar la recepta màster en funció de la demanda. Un darrer exemple defineix el procés de producció de fibra acrílica, il·lustrant decisions com la selecció de tasques, tecnologia, reactius o reutilització de dissolvents.
La industria productos químicos especiales se basa en la fabricación discontinua, la cual permite la adaptación frecuente de los sistemas de producción en función de las fluctuaciones de mercado. Para ser líder en el sector, son necesarias herramientas de soporte a la decisión que contribuyan al ágil desarrollo e implementación de nuevos procesos. Además, éstos deben ser competitivos para garantizar su viabilidad a largo plazo. Otras piezas clave para una operación eficiente son la utilización de plantas flexibles y el estudio de los fenómenos fisicoquímicos. Esta tesis aborda justamente el desarrollo sistemático de procesos químicos discontinuos que sean eficientes, económicamente competitivos y ecológicos, para contribuir a su rápida introducción en los sistemas de producción, ya sea en escenarios de plantas existentes o desde las bases. En particular, se plantea la resoluciónsimultánea de la síntesis conceptual de esquemas de proceso y la asignación de equipos, teniendo en cuenta además el diseño de planta.Con este fin, se propone una metodología de solución basada en optimización, donde todas las alternativas estructurales se representan en una Red de Estados y Equipos (SENpor sus siglas en inglés) que se formula mediante un problema de Optimización Dinámica Mixta-Lógica (MLDO por sus siglas en inglés) que se resuelve minimizando una función objetivo. La solidez de la metodología propuesta reside en la estrategia de modelado delproblema MLDO, que integra los diferentes tipos de decisiones en un solo modelo de optimización. En concreto, se consideran: (i) la combinación de alternativas de síntesis y asignación de equipos, (ii) modelos de proceso y trayectorias de control dinámicos, (iii)eventos discretos asociados al cambio de fase y operación, (iv) información cuantitativa y cualitativa, (v) sincronización de la transferencia de material en tareas consecutivas, y(vi) elementos de procesado discontinuos y semicontinuos.Existen diversas estrategias para resolver el problema MLDO resultante. En esta tesis se propone en primer lugar un método determinístico directo-simultáneo, donde el problema mixto-lógico se reformula en un mixto-entero. A su vez, éste se discretiza de formacompleta para obtener un problema de Programación No-Lineal Mixta-Entera (MINLP por sus siglas en inglés) el cual se puede resolver mediante algoritmos de optimización convencionales. Además, se presentan un Algoritmo Genético Diferencial (DGA por sussiglas en inglés) y un método híbrido. Ambas estrategias se plantean como alternativas de búsqueda con objeto de mantener la bondad de la solución y mejorar la eficacia de computación para tratar problemas de dimensión industrial.La metodología de solución propuesta se aplica al desarrollo de procesos discontinuos en escenarios con plantas existentes, teniendo en cuenta las restricciones físicas de los equipos. Un primer ejemplo aborda la fabricación de productos químicos basada en un sistema de reacciones competitivas. En concreto, se desarrolla y mejora el proceso de producción a implementar en una red de reactores considerando diferentes escenarios económicos, criterios de decisión, y modificaciones de planta. En un segundo ejemplo,se optimiza el proceso foto-Fenton a ser ejecutado en una planta piloto para eliminar contaminantes emergentes.Persiguiendo la integración del desarrollo de proceso con el diseño de plantas flexi-bles en escenarios base, se presenta asimismo una formulación estocástica en dos etapas para optimizar el beneficio esperado de acuerdo a varios escenarios de demanda. Paramanejar la complejidad de dicho problema se propone la utilización de una heurística.Como ejemplo, se plantea el diseño de una planta de base para implementar el anterior sistema de reacciones competitivas, donde decisiones como las trayectorias dinámicas de control o la configuración de equipos permiten adaptar la receta máster en función de lademandas. Por último, se presenta un ejemplo donde se define el proceso de producción de fibra acrílica, ilustrando decisiones como la selección de tareas, alternativas tecnológicas, reactivos químicos o la reutilización de disolventes.

Many important modern engineering problems involve satisfying multiple objectives. Simultaneous optimization of these objectives can be difficult as they compete for the same set of any given resources. One way to solve multiple-objective optimization is with the use of genetic algorithms (GA’s).

One can break down the structure of these multi-objective genetic algorithms (MOGA’s) into two different approaches. One approach is based on incorporating multiple objectives into a single fitness function which will evaluate how well a given solution solves the issue. The other approach uses multiple fitness functions, each representing a different objective, which when combined create a solution set of possible solutions to the problem. This project focuses on combining these approaches in order to make a hybrid model, which can benefit from combining the results of the previous two methods; incorporating a level of automation that allows for inference of a final solution based on different prioritization of each objective. This solution would not have been previously attainable by either standalone method.

This project is named the Automated Inferential Multi-Objective Optimization System (AIMOS), and it can be applied to a multitude of different problem types. In order to show its capabilities, AIMOS has been applied to a theoretical optimization problem used to measure the effectiveness of GA’s.

Technology scaling trends lead to shrinking of the individual elements like transistors and wires in digital systems. The main driving force behind this is cutting the cost of the systems while the systems are filled with extra functionalities. This is the reason why a 3 GHz Intel processor now is priced less than what a 50MHz processor was priced 10 years ago. As in most cases, this comes with a price. This price is the complex design process and problems that stem from the reduction in physical dimensions. As the transistors became smaller in size and the systems became faster, issues like power consumption, signal integrity, soft error tolerance, and testing became serious challenges. There is an increasing demand to put CAD tools in the design flow to address these issues at every step of the design process. First part of this research investigates circuit level techniques to reduce power consumption in digital systems. In second part, improving soft error tolerance of digital systems is considered as a trade off problem between power and reliability and a power aware dynamic soft error tolerance control strategy is developed. The objective of this research is to provide CAD tools and circuit design techniques to optimize power consumption and to increase soft error tolerance of digital circuits. Multiple supply and threshold voltages are used to reduce power consumption. Variable supply and threshold voltages are used together with variable capacitances to develop a dynamic soft error tolerance control scheme.

Over the course of 60 years, since the invention of the integrated circuit (IC), exponential improvements in cost, performance and power consumption were observed. Such advances have been strongly linked with the continuous reduction of the dimensions in manufactured ICs, but this trend has shown decreasing benefits as fundamental limits are reached. Notice that such tiny devices have increased variability, which generates unpredictable variations in the behavior of the manufactured devices. These uncertainties are typically addressed by defining margins on the clock period, estimated during the design phase. However, the overly conservative margins produce significant degradations in performance. Additionally, the evolution that enabled circuits with increasingly higher density of components, also resulted in an extremely complex IC design. At every step, electronic design automation (EDA) tools are challenged to handle this increasing complexity, requiring more powerful techniques to comply with the specification constraints within an affordable runtime. This thesis investigates alternatives in order to improve power, performance, area, and cost, using established IC manufacturing technologies. Advances in EDA are proposed in three distinct topics: area minimization using Boolean methods, area and delay reduction for designs based on fieldprogrammable gate array (FPGA), and an alternative clocking scheme to reduce timing margins. The first contribution consists of a technology-independent method for area minimization of combinational logic. Local optimization is applied on and-inverter graphs (AIGs), performing multi-output Boolean decomposition using two-literal divisors, targeting node count reduction. The second contribution regards two methods targeting technology mapping of FPGAs. On one hand, a functional decomposition approach, which uses the support size as cost function, exploring the inherent characteristics of FPGAs. On the other hand, an approach for recursive remapping, which reduces the structural bias of the subject graph, uses the mapping results as cost function, and obtains significant reductions in area and delay. The third contribution evaluates the dynamic variability mitigation and simplification of power delivery networks (PDNs) using an adaptive clocking scheme based on ring oscillator clocks (ROCs). The impact of the PDN parameters and ROC location is investigated, showing potential improvements in performance, leakage power and cost.
A lo largo de los 60 años desde la invención del circuito integrado (CI), se han producido mejoras exponenciales en su coste, rendimiento y consumo de energía. Ese progreso esta fuertemente vinculado a la reducción continua de las dimensiones de fabricación de los circuitos integrados, pero esta tendencia ha ido mostrando menos beneficios a medida que se alcanzan límites fundamentales. Estos dispositivos minúsculos tienen una mayor variabilidad, lo que genera variaciones impredecibles en el comportamiento de los dispositivos fabricados. Estas incertidumbres generalmente se abordan mediante la definición de márgenes en el período de reloj, estimado durante la fase de diseño. Sin embargo, márgenes excesivamente conservadores producen degradaciones significativas en el rendimiento. Además, la evolución que permitió crear circuitos con una densidad de componentes cada vez mayor, también incrementado la complejidad de los CI. En cada paso del proceso de diseño, las herramientas de automatización de diseño electrónico (EDA) se enfrentan al desafío de manejar esta complejidad creciente, lo que requiere técnicas más potentes para cumplir con las restricciones impuestas por las especificaciónes dentro de un tiempo de ejecución asequible. Esta tesis investiga alternativas para mejorar el consumo de energía, el rendimiento, el área y el coste, utilizando las tecnologías de fabricación de CI ya establecidas. Los avances en EDA se proponen en tres temas distintos: minimización de área usando métodos booleanos, reducción de retardo y área para diseños basados en matrices de puertas programables (FPGA), y un esquema de reloj alternativo para reducir los márgenes de tiempo excesivamente conservadores. La primera contribución consiste en un método independiente de tecnología para minimizar el área de la lógica combinacional. Se aplica optimización local en grafos de AND-inversor, utilizando una descomposición Booleana de múltiples salidas con divisores de dos literales, con el objetivo de reducir la cantidad de nodos. La segunda contribución propone dos métodos para el mapeo tecnológico de los FPGA. Por un lado, un método de descomposición funcional, que utiliza el tamaño del soporte como función de coste, explorando las características inherentes de los FPGA. Por otro lado, un método de mapeo recursivo, que reduce la distorsión estructural del grafo sujeto, utiliza los resultados del mapeo como función de coste y obtiene reducciones significativas en área y retardo. La tercera contribución evalúa la mitigación de la variabilidad dinámica y la simplificación de la red de suministro de energía (PDN) utilizando un esquema de reloj adaptativo basado en un oscilador en anillo (ROC). Se investiga el impacto de los parámetros de la PDN y la ubicación del ROC, mostrando mejoras potenciales en el rendimiento, consumo estático y coste.

Inductive logic programming (ILP) is a recently emerging subfield of machine learning that aims at overcoming the limitations of most attribute-value learning algorithms by adopting a more powerful language of first-order logic. Employing successful learning techniques of ILP to learn interesting characteristics among database relations is of particular interest to the knowledge discovery in databases research community. However, most existing ILP systems are general-purpose learners and that means users have to know how to tune some factors of ILP learners to best suit their tasks at hand. One such factor with great impact on the efficiency of ILP learning is how to specify the language bias. The language bias is a restriction on the format (or syntax) of clauses allowed in the hypothesis space. If the language is too weak, the search space is very large, and hence, the learning efficiency is decreased. On the contrary, if the language is too strong, the search space is so small that many interesting rules may be excluded from consideration. It is the purpose of this dissertation to develop an algorithm to generate a potentially useful language bias that is more appropriate for the task of inducing semantic constraints from the database relations. These constraints will be a major source of semantic knowledge for semantic query optimization in database query processing. The efficiency of the proposed algorithm was verified experimentally. The appropriate form of language bias specification, which is the output of the algorithm, was tested on the ILP system CLAUDIEN comparing with a number of different forms of language bias specification. The learning results were compared on the basis of number of rules discovered, the quality of rules, total time spent to learn rules, and the size of the search space. The experimental results showed that the proposed algorithm is helpful for the induction of semantic rules.

This thesis evaluates the benefit of meshing mathematical programming and expert systems for solving capital budgeting problems, using constraint logic programming methods. A review of modelling capabilities of mathematical programs for capital budgeting, and of financial expert systems leads to defining the respective role and potential of each method, and to the proposal of a two-tiered project selection approach: project evaluation and resource allocation. With emphasis placed on a tight coupling of the two tiers, logic programming is shown to be a language of choice to implement mathematical programming within an expert system shell. Prolog has the requisite properties to deal with both logical considerations and optimization problems. Although Prolog was not primarily designed to solve optimization problems, it is shown that the backtracking mechanism of the Prolog language is powerful enough for that purpose; it liberates the programmer from having to implement tree-search programs. A generate and test program is written in Turbo-Prolog, and compared to a more sophisticated test and generate implementation that uses methods of constraint satisfaction programming. Continuous capital budgeting problems are solved in CLP(${\cal R}$), an experimental extension of Prolog that enables the solution of simultaneous algebraic constraints, as required to solve linear programs.

Abstract Circulating Fluidized Bed (CFB) boilers are widely used for multi-fuel combustion of waste and bio-fuels. When several non-homogeneous fuels, having varying heat values, are burned simultaneously, the boiler control system can be affected by various control challenges, especially since it is not feasible to reliably measure the energy content of the multi-fuel flow. In order to fulfill energy production needs and maintain the ability to burn low grade fuels, co-firing with high heat value fuels such as gas, oil or coal is needed. Fuzzy Logic Control (FLC) has been successfully used for solving control challenges, where operators' process expertise can be transformed into automation. Real life control objects are often non-linear because the dynamics change with the operating point, or there might be other essential non-linearities in the combustion process. The proposed fuzzy control applications were developed to solve control challenges the operators meet in daily operation of a 150 MW(th) CFB at Varenso Oy's (Stora Enso Oyj) K6 boiler in Varkaus Finland. Before implementing the applications in the fullscale boiler, they were tested at a 2 MW(e) pilot plant boiler at Foster Wheeler Energia Oy's Research Center in Karhula, Finland. According to the industrial experiments, the four applications (steam pressure control, compensation of fuel quality fluctuation, fuel-feed optimization and increased bed inventory monitoring) discussed in this thesis, showed satisfactory performance and various improvements to the boiler control were achieved. Fuzzy logic control was shown to be a notable tool to improve the multi-fuel CFB boiler control.

Thesis (M. S.)--Industrial and Systems Engineering, Georgia Institute of Technology, 2008.
Committee Chair: Anton Kleywegt; Committee Co-Chair: Alexander Shapiro; Committee Member: Charles Rosa; Committee Member: Shabbir Ahmed; Committee Member: Sigrun Andradottir.

Thesis (Ph.D)--Industrial and Systems Engineering, Georgia Institute of Technology, 2009.
Committee Chair: Arkadi Nemirovski; Committee Co-Chair: Alexander Shapiro; Committee Co-Chair: Renato D. C. Monteiro; Committee Member: Anatoli Jouditski; Committee Member: Shabbir Ahmed. Part of the SMARTech Electronic Thesis and Dissertation Collection.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1994.
Includes bibliographical references (p. 91-93).
by Roxann Russell Blanchard.
M.S.

Il fenomeno dei Gamma Ray Burst (GRBs) scoperto casualmente nel 1967 dai satelliti Vela, si è dimostrato estremamente interessante per la cosmologia moderna e per le varie teorie di gravità quantistica. La correlazione dei GRBs con alcuni dei vari eventi capaci di emettere onde gravitazionali, permette di identificarli come controparti elettromagnetiche alla rivelazione di onde gravitazionali, aumentando la precisione nella localizzazioni delle sorgenti di queste ultime. Per sfruttare le peculiarità dei GRBs e approfondire la comprensione del fenomeno stesso è stato proposto il progetto High Energy Rapid Modular Ensemble of Satellites (HERMES). Questo progetto prevede il lancio di una costellazione di nanosatelliti, su differenti orbite, per la localizzazione dei transienti. L’obiettivo di questa tesi è quello di ottimizzare uno dei possibili algoritmi da implementare a bordo per la rivelazione di transienti. L'ottimizzazione è richiesta per massimizzare il ritorno scientifico della missione e per adattare l’algoritmo alle risorse disponibili a bordo di un nanosatellite. Tale ottimizzazione è stata eseguita simulando un data set di GRBs e applicandovi l’algoritmo. La migliore soluzione è stata identificata comparando l’efficienza di rivelazione ottenuta con varie combinazioni dei parametri dell’algoritmo.

Thesis (Ph. D.)--University of California, San Diego, 2007.
Title from first page of PDF file (viewed January 10, 2008). Available via ProQuest Digital Dissertations. Vita. Includes bibliographical references.

A new solution method for solving the real time production optimization (RTPO) problem for a petroleum production system is presented in this thesis. The objective function of the problem maximizes oil production and the RTPO handles decision variables at operational level. Including routing of production flows, lift gas allocation, and pressure configurations of the system. It is aimed to give decision support in a time horizon of days to weeks. Such problems require solution methods able to obtain solutions swiftly, as production planners adjust network components frequently to maintain optimal production.The problem contains binary decision variables combined with nonlinear expressions and is mathematically classified as a nonconvex mixed integer nonlinear problem (MINLP). MINLPs are in general known as computationally expensive and hard to solve to optimality, and when nonconvexities are present, few solvers can guarantee global optimality. The solution method presented deviates from traditional optimization techniques applied to such problems, and introduces logic disjunctions to substitute the binary variables of the MINLP. A specialized branch and bound algorithm (LBB) is developed to utilize the structure of these disjunctions, and as time is of paramount importance for the RTPO, it is aimed to reduce demanded computational effort for the problem. The LBB is given a high degree of user flexibility to be able to tailor the algorithm to different problems.Results of the LBB show substantial variation in solution efficiency when applied to a real petroleum production system. Only when specific problem knowledge is utilized to customize the algorithm to the current system, the algorithm provides solid reduction in computational effort compared to a recognized commercial solver. Also when applied to variations in system structure the LBB clearly outperforms the applied solver, and the effectiveness and robustness of the proposed algorithm when utilizing problem specific knowledge is confirmed. The fact that the LBB provides the same solution to the problem as the applied solver might also indicate that the nonconvexities of the problem are not as complex as expected, and that the solver is in fact able to find the global optimal solution.

Since integration technology is approaching the nanoelectronics range, some practical limits are being reached. Leakage power is increasing more and more with the continuous scaling, and design of clock distribution systems needs to be reconsidered as it becomes difficult to deal with performance and power consumption specifications while keeping a correct synchronisation in modern multi-GHz systems. The ongoing technology trend will become difficult to maintain unless dedicated library cells, new logic styles and circuit methods are emerging to prevent the drawbacks of future nanoscale circuits. In this thesis we investigate a new class of dynamic differential logic family that features a self-timed operation and low output logic swing. The latter contributes to reduce dynamic power, while the self-timing scheme alleviates the drawbacks of synchronous circuits and systems. Furthermore, the dynamic and differential nature of LSCML class brings advantages in terms of reduction of the power consumption variation and thus gives LSCML an additional potential for implementation of secure encryption devices against attacks based on power analysis. We investigate dynamic and leakage power reduction at the cell level through the application of low-power low-voltage techniques to a new hybrid full adder structure. The 8b RCA circuit based on the ULPFA (ultra low power full adder) version of this full adder, achieves a total power and a leakage power, which are both reduced by 50% compared to the 8b RCA implemented with conventional static CMOS full adder, while featuring better power delay product.

This thesis describes the development and experimental results of a system to explore cloning of facial expressions between dissimilar face models, so new faces can be animated using the animations from existing faces. The system described in this thesis uses fuzzy membership functions and subtractive clustering to represent faces and expressions in an intermediate space. This intermediate space allows expressions for face models with different resolutions to be compared. The algorithm is trained for each pair of faces using particle swarm optimization, which selects appropriate weights and radii to construct the intermediate space. These techniques allow the described system to be more flexible than previous systems, since it does not require prior knowledge of the underlying implementation of the face models to function.

The widespread adoption of embedded computing systems has resulted in the realization of numerous sensing, decision, and control applications with diverse application-specific requirements. However, such embedded systems applications are becoming increasingly difficult to design, simulate, and optimize due to the multitude of interdependent parameters that must be considered to achieve optimal, or near-optimal, performance that meets design constraints. This situation is further exacerbated for data-adaptable embedded systems (DAES) applications due to the dynamic characteristics of the deployment environment and the data streams on which these systems operate. As operating conditions change, these embedded systems must continue to adapt their configuration and composition at runtime in order to meet application requirements. To assist both platform developers and application domain experts, this dissertation presents design and optimization frameworks for the synthesis of runtime adaptable embedded systems. For sensor network applications, we present an initial dynamic profiling and optimization platform that profiles network and sensor node activity to generate optimal node configurations at runtime based on designed-specified application requirements. To support a broader class of DAES applications, we present a modeling and optimization framework that supports the specification of application task flows, data types, and runtime estimation models for the runtime adaptation of task implementations and device mappings. Experimental results for these design and optimization frameworks demonstrate the benefits of dynamic optimization compared to static optimization alternatives. For the presented sensor network and video-based collision avoidance applications, dynamic configurations exhibited improvements of up to 109% and 76%, respectively. Moreover, the performance of the heuristic design space exploration (DSE) algorithms utilized by the runtime optimization frameworks is compared to exhaustive DSE implementations, resulting in speedups of up to 1662X and 544X for the same two applications, respectively.

This research is focused on helping engineers design better systems by supporting their decision making. When engineers design a system, they have an almost unlimited number of possible system alternatives to consider. Modern systems are difficult to design because of a need to satisfy many different stakeholder concerns from a number of domains which requires a large amount of expert knowledge. Current systems engineering practices try to simplify the design process by providing practical approaches to managing the large amount of knowledge and information needed during the process. Although these methods make designing a system more practical, they do not support a structured decision making process, especially at early stages when designers are selecting the appropriate system architecture, and instead rely on designers using ad hoc frameworks that are often self-contradictory. In this dissertation, a framework for performing architecture exploration at early stages of the design process is presented. The goal is to support more rational and self-consistent decision making by allowing designers to explicitly represent their architecture exploration problem and then use computational tools to perform this exploration. To represent the architecture exploration problem, a modeling language is presented which explicitly models the problem as an architecture selection decision. This language is based on the principles of decision-based design and decision theory, where decisions are made by picking the alternative that results in the most preferred expected outcome. The language is designed to capture potential alternatives in a compact form, analysis knowledge used to predict the quality of a particular alternative, and evaluation criteria to differentiate and rank outcomes. This language is based on the Object Management Group's System Modeling Language (SysML). Where possible, existing SysML constructs are used; when additional constructs are needed, SysML's profile mechanism is used to extend the language. Simply modeling the selection decision explicitly is not sufficient, computational tools are also needed to explore the space of possible solutions and inform designers about the selection of the appropriate alternative. In this investigation, computational tools from the mathematical programming domain are considered for this purpose. A framework for modeling an architecture selection decision in mixed-integer linear programming (MIP) is presented. MIP solvers can then solve the MIP problem to identify promising candidate architectures at early stages of the design process. Mathematical programming is a common optimization domain, but it is rarely used in this context because of the difficulty of manually formulating an architecture selection or exploration problem as a mathematical programming optimization problem. The formulation is presented in a modular fashion; this enables the definition of a model transformation that can be applied to transform the more compact SysML representation into the mathematical programming problem, which is also presented. A modular superstructure representation is used to model the design space; in a superstructure a union of all potential architectures is represented as a set of discrete and continuous variables. Algebraic constraints are added to describe both acceptable variable combinations and system behavior to allow the solver to eliminate clearly poor alternatives and identify promising alternatives. The overall framework is demonstrated on the selection of an actuation subsystem for a hydraulic excavator. This example is chosen because of the variety of potential architecture embodiments and also a plethora of well-known configurations which can be used to verify the results.