Mimicking natural evolution to solve hard optimization problems has played an important role in the artificial intelligence arena. Such techniques are broadly classified as Evolutionary Algorithms (EAs) and have been investigated for around four decades during which important contributions and advances have been made. One main evolutionary technique which has been widely investigated is the Genetic Algorithm (GA). GAs are stochastic search techniques that follow the Darwinian principle of evolution. Their application in the solution of hard optimization problems has been very successful. Indeed multi-dimensional problems presenting difficult search spaces with characteristics such as multi-modality, epistasis, non regularity, deceptiveness, etc., have all been effectively tackled by GAs. In this research, a competitive form of GAs known as fine or cellular GAs (cGAs) are investigated, because of their suitability for System on Chip (SoC) implementation when tackling real-time problems. Cellular GAs have also attracted the attention of researchers due to their high performance, ease of implementation and massive parallelism. In addition, cGAs inherently possess a number of structural configuration parameters which make them capable of sustaining diversity during evolution and therefore of promoting an adequate balance between exploitative and explorative stages of the search. The fast technological development of Integrated Circuits (ICs) has allowed a considerable increase in compactness and therefore in density. As a result, it is nowadays possible to have millions of gates and transistor based circuits in very small silicon areas. Operational complexity has also significantly increased and consequently other setbacks have emerged, such as the presence of faults that commonly appear in the form of single or multiple bit flips. Tough environmental or time dependent operating conditions can trigger faults in registers and memory allocations due to induced radiation, electron migration and dielectric breakdown. These kinds of faults are known as Single Event Effects (SEEs). Research has shown that an effective way of dealing with SEEs consists of a combination of hardware and software mitigation techniques to overcome faulty scenarios. Permanent faults known as Single Hard Errors (SHEs) and temporary faults known as Single Event Upsets (SEUs) are common SEEs. This thesis aims to investigate the inherent abilities of cellular GAs to deal with SHEs and SEUs at algorithmic level. A hard real-time application is targeted: calculating the attitude parameters for navigation in vehicles using Global Positioning System (GPS) technology. Faulty critical data, which can cause a system’s functionality to fail, are evaluated. The proposed mitigation techniques show cGAs ability to deal with up to 40% stuck at zero and 30% stuck at one faults in chromosomes bits and fitness score cells. Due to the non-deterministic nature of GAs, dynamic on-the-fly algorithmic and parametric configuration has also attracted the attention of researchers. In this respect, the structural properties of cellular GAs provide a valuable attribute to influence their selection pressure. This helps to maintain an adequate exploitation-exploration tradeoff, either from a pure topological perspective or through genetic operations that also make use of structural characteristics in cGAs. These properties, unique to cGAs, are further investigated in this thesis through a set of middle to high difficulty benchmark problems. Experimental results show that the proposed dynamic techniques enhance the overall performance of cGAs in most benchmark problems. Finally, being structurally attached, the dimensionality of cellular GAs is another line of investigation. 1D and 2D structures have normally been used to test cGAs at algorithm and implementation levels. Although 3D-cGAs are an immediate extension, not enough attention has been paid to them, and so a comparative study on the dimensionality of cGAs is carried out. Having shorter radii, 3D-cGAs present a faster dissemination of solutions and have denser neighbourhoods. Empirical results reported in this thesis show that 3D-cGAs achieve better efficiency when solving multi-modal and epistatic problems. In future, the performance improvements of 3D-cGAs will merge with the latest benefits that 3D integration technology has demonstrated, such as reductions in routing length, in interconnection delays and in power consumption.

Approved for public release; distribution is unlimited / This thesis considers the problem of generating optimal entry trajectories for a reusable launch vehicle following a control surface failure. The thesis builds upon the work of Dr. David Doman, Dr. Michael Oppenheimer and Dr. Michael Bolender of the Air Vehicles Directorate, Air Force Research Lab Dayton Ohio. The primary focus of this work is to demonstrate the feasibility of inner loop reconfiguration and outer loop trajectory retargeting and replanning for the X-33 reusable launch vehicle (RLV) following the imposition of a control surface failure. The trajectory generation model employs path constraints generated by an AFRL trim deficiency algorithm coupled with an inner loop control allocator and aerodynamic database that captures the full 6-DOF vehicle aerodynamic effects while utilizing an outer loop 3-DOF model. The resulting optimal trajectory does not violate the trim deficiency constraints and provides additional margins for trajectories flown during failure conditions. The footprints generated by the thesis show that contemporary footprint analysis for vehicles experiencing control surface failures are overly optimistic when compared to those footprints that consider vehicle aerodynamic stability and realistic landable attitudes at the threshold of the landing runway. The results of the thesis also show the performance reductions resulting from decoupling the inner and outer loop and that trajectories can be generated to the landing runway without using a region of terminal area energy management. / Commander, United States Navy

Approved for public release, distribution is unlimited / Numerous optimization problems involve systems with multiple and often contradictory criteria. Such contradictory criteria have been an issue for marine/naval engineering design studies for many years. This problem becomes more important when one considers novel ship types with very limited or no operational record. A number of approaches have been proposed to overcome these multiple criteria design optimization problems. This Thesis follows the Parameter Space Investigation (PSI) technique to address these problems. The PSI method is implemented with a software package called MOVI (Multi-criteria Optimization and Vector Identification). Two marine/naval engineering design optimization models were investigated using the PSI technique along with the MOVI software. The first example was a bulk carrier design model which was previously studied with other optimization methods. This model, which was selected due to its relatively small dimensionality and the availability of existing studies, was utilized in order to demonstrate and validate the features of the proposed approach. A more realistic example was based on the "MIT Functional Ship Design Synthesis Model" with a greater number of parameters, criteria, and functional constraints. A series of optimization studies conducted for this model demonstrated that the proposed approach can be implemented in a naval ship design environment and can lead to a large design parameter space exploration with minimum computational effort. / Lieutenant Junior Grade, Turkish Navy

In multi-tier cellular networks, user performance is largely a ected by the varying transmit powers, distances, and non-uniform tra c loads of di erent base stations (BSs) in both the downlink (DL) and uplink (UL) directions of transmission. In presence of such heterogeneity, decoupled UL-DL user association (DUDe), which allows users to associate with di erent BSs for UL and DL transmissions, can be used to optimize network performance. Again, in-band full-duplex (FD) communi- cation is considered as a promising technique to improve the spectral e ciency of future multi-tier fth generation (5G) cellular networks. Nonetheless, due to severe UL-to-DL and DL-to-UL interference issues arising due to FD communications, the performance gains of DUDe in FD multi-tier networks are inconspicuous. To this end, this thesis develops a comprehensive framework to analyze the usefulness of DUDe in a full-duplex multi-tier cellular network. We rst formulate a joint UL and DL user association problem (with the provision of decoupled association) that maximizes the sum-rate for UL and DL transmission of all users. Since the formulated problem is a mixed-integer non-linear programming (MINLP) problem, we invoke approxi- mations and binary constraint relaxations to convert the problem into a Geometric Programming (GP) problem that is solved using Karush-Kuhn-Tucker (KKT) opti- mality conditions. Given the centralized nature and complexity of the GP problem, the solution of which serves as the upper bound for any sub-optimal solution, we formulate a distributed two-sided iterative matching game and develop a solution to obtain the solution of the game. In this game, the users and BSs rank one another using preference metrics that are subject to the externalities (i.e., dynamic interfer- ence conditions). The solution of the game is guaranteed to converge and provides Pareto-e cient stable associations. Finally, we derive e cient light-weight versions of the iterative matching solution, i.e., non-iterative matching and sequential UL-DL matching algorithms. The performances of all the solutions are critically evaluated in terms of aggregate UL and DL rates of all users, the number of unassociated users, and the number of coupled/decoupled associations. Simulation results demonstrate the e cacy of the proposed algorithms over the centralized GP solution as well as traditional coupled and decoupled user association schemes. / October 2016

<p>U disertaciji se prikazuje razvoj sistema za<br />projektovanje i optimizaciju konstrukcije pribora za<br />ma&scaron;insku obradu. Analizirani su različiti prilazi u<br />projektovanu pribora. Prikazana je koncepcija i<br />funkcionisanje razvijenog sistema. Validacija sistema<br />je izvr&scaron;ena na konkternim radnim predmetima za<br />operacije obrade bu&scaron;enja i glodanja. Na kraju su dati<br />odgovarajući zaključci i mogući pravci budućih<br />istraživanja</p> / <p> The dissertation shows the development of<br /> system for fixture design and layout<br /> optimization for machining processi. Different<br /> approaches were analyzed in fixture design.<br /> The concept and functioning of the developed<br /> system is presented. Validation of the system<br /> is made with specific workpieces for drilling<br /> and milling process operations. At the end the<br /> appropriate conclusions and possible<br /> directions for future research are given.</p>

<p>U doktorskoj disertaciji je prikazana nova metodologija za projektovanje<br />i optimizaciju konstrukcije elemenata pribora. Projektovani su i<br />realizovani uređaji koji omogućavaju ispitivanje nosivosti i<br />popustljivosti kontakta između elemenata pribora i radnog predmeta u<br />statičkim i dinamičkim uslovima opterećenja. U istraživanjima je<br />simuliran proces stezanja elementima sa specijalno projektovanim<br />završetkom i praćena je nosivost i popustljivost spoja između<br />elemenata pribora i radnog predmeta. Utvrđeno je da standardni<br />elementi za stezanje sa ravnim čelom u odnosu na specijano<br />projektovane elemente imaju značajno manju nosivost i popustljivost.<br />Pozitivni efekti primene elemenata za stezanje sa specijalno<br />projektovanim završetkom ogledaju se u povećanju pouzdanosti, tačnosti<br />i produktivnosti mašinske obrade.</p> / <p>Presented in this doctoral dissertation is a new methodology for the design and<br />optimization of fixture elements. Special device is designed and manufactured<br />to test load capacity and interface compliance between fixture elements and<br />workpiece under static and dynamic loads during machining. The research<br />process is simulated by specially designed clamping elements and monitored<br />for load capacity and interface compliance between fixture elements and<br />workpiece. It was found that the standard clamping elements with flat clamping<br />surface have a significantly lower load capacity and interface compliance in<br />comparison with the specially designed clamping elements. Application of the<br />specially designed clamping elements results in increased reliability, accuracy<br />and machining productivity.</p>

The focus of the present master thesis is the automation of an existing controllerdesign for a missile using two aerodynamic actuating systems. The motivation isto evaluate more missile concepts in a shorter period of time.The option used is trimming and linearization of a highly nonlinear missile at specic conditions. According to these conditions, either a two-dimensional operatingpoint grid dened by Mach number and height or three-dimensional operatingpoint grid dened by Mach number, height and angle of attack is generated forthe whole operating range of the missile. The controllers are designed at thesepoints using convex optimization. The convex set denes the pole placement areawhich is constrained by linear matrix inequalities according to the dynamic behaviorof the missile at the operating point conditions. These controllers describea validity area where the missile can be stabilized. This area consists all neighboringoperating points and denes therefore the grid density which can dier atspecic regions of the operating range. Controlling the missile to the target makesit necessary to apply gain-scheduling in order to get the manipulated variable byinterpolation of adjacent operating points. During this blending of the controllersa problem called windup can occur when an actuator is saturated. This mightlead to instability in worst case but can be counteracted by a model-recovery antiwindupnetwork which guarantees stability in the presence of saturation. Thisanti-windup design is automated by an ane linear parameter dependency of thegrid parameters and has the same validity area like the controllers.The whole design was successfully developed and tested in MATLAB/Simulink onmissiles using one or two aerodynamic actuating systems. The controllers have agood performance at small and high acceleration steps and the anti-windup keepsthe missile stable even though the actuators are saturated. Stability and robustnessof the controllers and anti-windup networks was veried as well as an airdefense maneuver where the missile starts at the ground and intercepts a targetat high altitude was successfully simulated for dierent grids and missiles.

Context: Recent trends show that there is a tremendous shift from IT companies following traditional methods by hosting their applications/systems in self-managed on premise data centers to using the so-called cloud data centers. Cloud computing has received immense popularity due to its architecture and the ease of usage. Due to this increase in demand and shift in practices, there has been a tremendous increase in number of data centers over a period, resulting in increase of energy consumption. In this thesis work, a research is carried out on optimizing the energy consumption of a typical cloud data center. OpenStack cloud computing software is chosen as the platform in this research. We have used live migration as a key aspect in this research. Objectives: In this research, our objectives are as follows: Design an OpenStack testbed to implement the migration of virtual machines. To estimate the energy consumption of the data center. To design a heuristic algorithm to evaluate the performance metrics and to optimize the overall energy consumption. Methods: We have used PowerAPI, a software tool to estimate the energy consumption of hosts as well as virtual machines. A heuristic algorithm is designed and implemented in an instrumental OpenStack testbed to optimize the energy consumption. Server consolidation and load balancing of virtual machines methodologies are used in the heuristic algorithm design. Our research is carried out against the functionality of Nova scheduler of OpenStack. Results: Results section describes the values of performance metrics yielded by carrying out the experiment. The obtained results showed that energy can be optimized significantly by modifying the way OpenStack nova scheduler can work. The experiment is carried out on vanilla OpenStack and OpenStack with the heuristic algorithm in place, In the second case, the nova scheduler algorithms are not used but the heuristic algorithm is used instead. The CPU utilization and CPU load were noticed to be higher than the metrics observed in case of OpenStack with nova scheduler. Energy consumption is observed to be lesser than the consumption in OpenStack design with nova scheduler. Conclusions: The research tells that energy consumption can be optimized significantly using desired algorithms without compromising the service quality it offers. However, the design impacts on CPU slightly as the metrics are observed to be higher when compared to that in case of OpenStack with nova scheduler. Although it won’t have noticeable impact on the system.

Es erfolgt eine Darstellung der theoretischen Grundlagen, die für die Berechnung von optimal wachsenden Störungen nötig sind. Bei Anwendung des Projektionsoperators ist es möglich, die Anfangsstörung zu berechnen, welche die größtmögliche kinetische Energie nach Ablauf eines Optimierungsintervalls in einem vorher spezifizierten Gebiet erlangt. Die Berechnung dieser sogenannten lokalen optimalen Vektoren erfolgt für ein barotropes Modell, das um zeitunabhängige Grundströme linearisiert ist. Die lokalen optimalen Vektoren sind für Grundströme ermittelt worden, die aus einem gleitenden 10-Tages Mittel bestehen. Als Zielgebiet kam dabei der Nordatlantik-Europa Bereich zur Anwendung. Auf diese Weise war es möglich, zu untersuchen, welche Gebiete einen Einfluss auf das Störungswachstum besitzen. Dabei stellte sich heraus, dass für Optimierungszeiten größer 96 Stunden vier verschiedene Startgebiete existieren. Besonderes Interesse galt dem Transport von kinetischer Energie aus diesen Startgebieten in das Zielgebiet. Teilweise zeigte sich, dass die Energie über PNA ähnliche Strukturen in das Zielgebiet getragen wird. / Under the use of a projection operator it is possible to calculate perturbations with maximal energy growth during optimization time at some pre-chosen geographical area. The so called local optimal perturbations are computed for a barotropic model, linarized about time independent basic states. The optimal perturbations are determined for basic states consisting of 10-day running means of the daily streamfunctions. The chosen target area was the North Atlantic-European sector. In this way it was possible to find relevant \"starting regions\" leading to optimal energy growth in the target area. We found four such regions for long optimization times (> 96 hours). Thereby, the mechanisms responsible for the transport of the perturbations from the starting regions into the target area was thereby of special interest. In particular cases we found that structures similar to Pacific North America (PNA) patterns are involved in the energy transfer.

The thesis presents models for short term planning (24 hours) of a hyro dominated hydrothermal power system. The purpose of the models is to minimizae the system operation costs to provide a forecasted load and keep enough spinning reserve. / This thesis presents models for short term planning (24 hours) of a hydro dominated hydrothermal power system. The purpose of the models is to minimize the system operation cost to provide a forecasted load and keep enough spinning reserve.   The thesis focuses on two issues in hydro power modelling. The first issue is the relationship between water discharged and power generated. This relationship is a non-linear and non-convex function. If the plant has several units, the efficiency of the plant will have local maximums, so called local best-efficiency points. The second issue is to take into account the cost of start-ups of hydro units in the planning.   The hydro model is mixed-integer. Discharg􀁐s are allowed at zero flow, the local best-efficiency points and on the continuous part between the local best-efficiency point with the highest flow and the point with maximum flow. This last continuous part is modelled as a linear function. In order to get data for the start-up cost a survey among the largest power producers in Sweden has been made, where three questions about start-ups of hydro power units has been asked: What causes the costs in the start-up?, How much does a start-up cost? and How do start-ups effect the short-term scheduling strategies of power producers in Sweden? The results show that a fair estimate of the start-up cost is about $3/MW nominal output. For the thermal plants a standard model with polynomial operation cost, start-up costs and ramp-rate constraints has been used. The model also includes the possibilities of purchasing and selling power to forecasted prices.   The planning problem is formulated as a mathematical programming problem. The solution technique uses Lagrange relaxation to decompose the problem into subproblems. There will be one subproblem for each hydro and thermal plant. In order to find good feasible solutions a heuristic technique to change the integer variables in the hydro system has been developed. The Lagrange multipliers are updated with the subgradient method.   The models are tested in three different load situations; a winter day (heavy load), an autumn day (medium load) and a summer day (light load). The result shows that the method gives near optimal schedules in reasonable computation time in cases with a normal part of the thermal units committed. The assumed start-up cost results in that hydro units almost never are started or stopped for one hour only. / <p>QC 20161206</p>