Thesis (MComm)--Stellenbosch University, 2001. / ENGLISH ABSTRACT: This thesis describes the design and implementation of a data storage service. Many clients can be served simultaneously in an environment where processes execute on different physical machines and communicate via message passing primitives. The service is provided by two separate servers: one that functions at the disk block level and another that maintains files. A prototype system was developed first in the form of a simple file store. The prototype served two purposes: (1) it extended the single-user Oberon system to create a multiuser system suitable to support group work in laboratories, and (2) it provided a system that could be measured to obtain useful data to design the final system. Clients access the service from Oberon workstations. The Oberon file system (known as the Ceres file system) normally stores files on a local disk. This system was modified to store files on a remote Unix machine. Heavily used files are cached to improve the efficiency of the system. In the final version of the system disk blocks are cached, not entire files. In this way the disks used to store the data are unified and presented as a separate virtual block service to be used by file systems running on client workstations. The virtual block server runs on a separate machine and is accessed via a network. The simplicity of the block server is appealing and should in itself improve reliability. The main concern is efficiency and the goal of the project was to determine whether such a design can be made efficient enough to serve its purpose. / AFRIKAANSE OPSOMMING:Hierdie tesis omskryf die ontwerp en implementasie van 'n data stoor diens. Verskeie gebruikers word bedien deur die diens wat funksioneer in 'n verspreide omgewing: 'n omgewing waar prosesse uitvoer op verskillende masjiene en met mekaar kommunikeer met behulp van boodskappe wat rondgestuur word. Die diens word verskaf deur twee bedieners: die eerste wat funksioneer op 'n blok vlak en die ander wat lers onderhou. 'n Prototipe leer diens is ontwikkel deur middel van 'n basiese leer stoor. Die prototipe het twee funksies verrig: (1) die enkel gebruiker Oberon stelsel is uitgebrei na 'n veelvoudige gebruiker stelsel bruikbaar vir groepwerk in 'n laboratorium omgewing, en (2) 'n stelsel is verskaf wat betroubare en akkurate data kon verskaf vir die ontwerp van die finale stelsel. Oberon werkstasies word gebruik met die leer diens. Die Oberon leer stelsel (ook bekend as die Ceres leer stelsel) stoor normaalweg leers op 'n lokale skyf. Hierdie bestaande stelsel is verander om leers te stoor op 'n eksterne Unix masjien. Leers wat die meeste in gebruik is word in geheue aangehou vir effektiwiteits redes. Die finale weergawe van die stelsel berg skyf blokke in geheue, nie leers nie. Hierdie metode laat dit toe om data te stoor op 'n standaard metode, bruikbaar deur verskillende tipes leer stelsels wat uitvoer op verskeie gebruikers se werkstasies. Die virtuele blok stoor voer uit op 'n aparte masjien en is bereikbaar via 'n netwerk. Die eenvoudige ontwerp van die diens is opsigself aanloklik en behoort betroubaarheid te verbeter. Die hoof bekommernis is effektiwiteit en die hoofdoel van die projek was om te bepaal of hierdie ontwerp effektief genoeg gemaak kon word.

Thesis (MSc)--University of Stellenbosch, 2001. / ENGLISH ABSTRACT: lP networks offer scalability and flexibility for rapid deployment of value added lP services. However, with the increased demand and explosive growth of the Internet, carriers require a network infrastructure that is dependable, predictable, and offers consistent network performance. This thesis examines the functionality, performance and implementation aspects of the MPLS mechanisms to minimize the expected packet delay in MPLS networks. Optimal path selection and the assignment of bandwidth to those paths for minimizing the average packet delay are investigated. We present an efficient flow deviation algorithm (EFDA) which assigns a small amount of flow from a set of routes connecting each OD pair to the shortest path connecting the OD pair in the network. The flow is assigned in such a way that the network average packet delay is minimized. Bellman's algorithm is used to find the shortest routes between all OD pairs. The thesis studies the problem of determining the routes between an OD pair and assigning capacities to those routes. The EFDA algorithm iteratively determines the global minimum of the objective function. We also use the optimal flows to compute the optimal link capacities in both single and multirate networks. The algorithm has been applied to several examples and to different models of networks. The results are used to evaluate the performance of the EFDA algorithm and compare the optimal solutions obtained with different starting topologies and different techniques. They all fall within a close cost-performance range. They are all within the same range from the optimal solution as well. / AFRIKAANSE OPSOMMING: lP-netwerke voorsien die skaleerbaarheid en buigsaamheid vir die vinnige ontplooing van toegevoegde-waarde lP-dienste. Die vergrote aanvraag en eksplosiewe uitbreiding van die Internet benodig betroubare, voorspelbare en bestendige netwerkprestasie. Hierdie tesis ondersoek die funksionaliteit, prestasie en implementering van die MPLS(multiprotokoletiketskakel)- meganismes om die verwagte pakketvertraging te minimeer. Ons bespreek 'n doeltreffende algoritme vir vloei-afwyking (EFDA) wat 'n klein hoeveelheid vloei toewys uit die versameling van roetes wat elke OT(oorsprong-teiken)- paar verbind aan die kortste pad wat die OT-paar koppel. Die vloei word toegewys sodanig dat die netwerk se gemiddelde pakketvertraging geminimeer word. Bellman se algoritme word gebruik om die kortste roetes tussen alle OT-pare te bepaal. Die tesis bespreek die probleem van die bepaling van roetes tussen 'n OT-paar en die toewysing van kapasiteite aan sulke roetes. Die EFDA-algoritme bepaal die globale minimum iteratief. Ons gebruik ook optimale vloeie vir die berekening van die optimale skakelkapasiteite in beide enkel- en multikoers netwerke. Die algoritme is toegepas op verskeie voorbeelde en op verskillende netwerkmodelle. Die skakelkapasiteite word aangewend om die prestasie van die EFDAalgoritme te evalueer en dit te vergelyk met die optimale oplossings verkry met verskillende aanvangstopologieë en tegnieke. Die resultate val binne klein koste-prestasie perke wat ook na aan die optimale oplossing lê.

Thesis (MSc)--University of Stellenbosch, 2002. / ENGLISH ABSTRACT: The science of signature verification is concerned with identifying individuals by their handwritten signatures. It is assumed that the signature as such is a unique feature amongst individuals and the creation thereof requires a substantial amount of hidden information which makes it difficult for another individual to reproduce the signature. Modern technology has produced devices which are able to capture information about the signing process beyond what is visible to the naked eye. A dynamic signature verification system is concerned with utilizing not only visible, i.e. shape related information but also invisible, hidden dynamical characteristics of signatures. These signature characteristics need to be subjected to analysis and modelling in order to automate use of signatures as an identification metric. We investigate the applicability of hidden Markov models to the problem of modelling signature characteristics and test their ability to distinguish between authentic signatures and forgeries. / AFRIKAANSE OPSOMMING: Die wetenskap van handtekeningverifikasie is gemoeid met die identifisering van individue deur gebruik te maak van hulle persoonlike handtekening. Dit berus op die aanname dat 'n handtekening as sulks uniek is tot elke individu en die generering daarvan 'n genoeg mate van verskuilde inligting bevat om die duplisering daarvan moeilik te maak vir 'n ander individu. Moderne tegnologie het toestelle tevoorskyn gebring wat die opname van eienskappe van die handtekeningproses buite die bestek van visuele waarneming moontlik maak. Dinamiese handtekeningverifikasie is gemoeid met die gebruik nie alleen van die sigbare manefestering van 'n handtekening nie, maar ook van die verskuilde dinamiese inligting daarvan om dit sodoende 'n lewensvatbare tegniek vir die identifikasie van individue te maak. Hierdie sigbare en onsigbare eienskappe moet aan analise en modellering onderwerp word in die proses van outomatisering van persoonidentifikasie deur handtekeninge. Ons ondersoek die toepasbaarheid van verskuilde Markov-modelle tot die modelleringsprobleem van handtekeningkarakteristieke en toets die vermoë daarvan om te onderskei tussen egte en vervalste handtekeninge.

Thesis (MSc)--Stellenbosch University , 1993. / ENGLISH ABSTRACT: Reactive systems are event driven state machines which usually do not terminate, but remain in perpetual interaction with their environment. Such systems usually interact 'With devices which introduce a high degree of concurrency and some real time constraints to the system. Because of the concurrent nature of reactive systems they are commonly implemented as communicating concurrent processes on one or more processors. Jeffay introduces a design paradigm which requires consumer processes to consume messages faster than they are produced by producer processes. If this is guaranteed, the real time constraints of such .. system are always met, and the correctness of the process interaction is guaranteed in terms of the message passing semantics. I developed the ESE kernel, which supports Jeffay systems by providing lightweight processes which communicate over asynchronous channels. Processes are scheduled non-preemptively according to the earliest deadline first policy when they have messages pending on their input channels. The Jeffay design method and the ESE kernel have been found to be highly suitable to implement embedded reactive systems. The general requirements of embedded reactive systems, and kernel support required by such systems, are discussed. / AFRIKAANSE OPSOMMING: Reaktiewe stelsels is toeatandsoutomate wat aangedryf word deur gebeure in hul omgewins. So 'n stelsel termineer gewoonlik nie, maar bly in 'n voortdurende wisselwerking met toestelle in sy omgewing. Toestelle in die omgewing van 'n reaktiewe stelsel veroorsaak in die algemeen 'n hoë mate van gelyklopendheid in die stelsel, en plaas gewoonlik sekere intydse beperkings op die stelsel. Gelyklopende stelsels word gewoonlik as stelsel. van kommunikerende prosesse geïmplementeer op een of meer prosessors. Jeffay beskryf 'n ontwerpsmetodologie waarvolgens die ontvanger van boodskappe hulle vinniger moet verwerk as wat die sender hulle kan stuur. Indien hierdie gedrag tussen alle pare kommunikerende prosesse gewaarborg kan word, sal die stelsel altyd sy intydse beperkings gehoorsaam, en word die korrektheid van interaksies tussen prosesse deur die semantiek van die boodskapwisseling gewaarborg. Die "ESE" bedryfstelselkern wat ek ontwikkel het, ondersteun stelsels wat ontwerp en geïmplementeer word volgens Jeffay se metode. Prosesse kommunikeer oor asinkrone kanale, en die ontvanger van die boodskap met die vroegste keertyd word altyd eerste geskeduleer. Jeffay se ontwerpsmetode en die "ESE" kern blyk in die praktyk baie geskik te wees vir reaktiewe stelsels wat as substelsels van groter stelsels uitvoer. Die vereistes van reaktiewe substelsels, en die kemondersteuning wat daarvoor nodig is, word bespreek.

Thesis (MSc)--Stellenbosch University, 1993. / ENGLISH ABSTRACT: Formal methods aim to apply the rigour of mathematical logic to the problem ofguaranteeing that the behaviour of (critical) software conforms to predetermined requirements. The application of formal methods during program construction centers around a formal specification of the required behaviour of the program. A development attempt is successful if the resulting program can be formally proven to conform to its specification. For any substantial program, this entails a great deal of effort. Thus, some research efforts have been directed at providing mechanical support for the application of formal methods to software development. E.W. Dijkstra's calculus of weakest precondition predicate transformers [39,38] represents one of the first attempts to use program correctness requirements to guide program development in a formal manner. / AFRIKAANSE OPSOMMING: Formele metodes poog om die strengheid van wiskundige logika te gebruik om te waarborg dat die gedrag van (kritiese) programmatuur voldoen aan gegewe vereistes. Die toepassing van formele metodes tydens programontwikkeling sentreer rondom a formele spesifikasie van die verlangde programgedrag. 'n Ontwikkelingspoging is suksesvol as daar formee1 bewys kan word dat die resulterende program aan sy spesifikasie voldoen. Vir enige substansiële program, verteenwoordig dit ‘n aansienlike hoeveelheid werk. Verskeie navorsinspoging is gerig op die daarstelling van meganiese ondersteuning vir die gebruik van formele metodes tydens ontwikkeling van sagteware. E. W. Dijkstra se calculus van swakste voorkondisie (“weakest precondition”) predikaattransformators [39,38] is een van die eerste pogings om vereistes vir programkorrektheid op ‘n formele en konstruktiewe wyse tydens programontwikkeling te gebruik.

Thesis (MSc)--Stellenbosch University, 2014. / ENGLISH ABSTRACT: In computer Go, moves are typically selected with the aid of a tree search algorithm. Monte-Carlo tree search (MCTS) is currently the dominant algorithm in computer Go. It has been shown that the inclusion of domain knowledge in MCTS is able to vastly improve the strength of MCTS engines. A successful approach to representing domain knowledge in computer Go is the use of appropriately weighted tactical features and pattern features, which are comprised of a number of hand-crafted heuristics and a collection of patterns respectively. However, tactical features are hand-crafted specifically for Go, and pattern features are Go-specific, making it unclear how they can be easily transferred to other domains. As such, this work proposes a new approach to representing domain knowledge, decision tree features. These features evaluate a state-action pair by descending a decision tree, with queries recursively partitioning the state-action pair input space, and returning a weight corresponding to the partition element represented by the resultant leaf node. In this work, decision tree features are applied to computer Go, in order to determine their feasibility in comparison to state-of-the-art use of tactical and pattern features. In this application of decision tree features, each query in the decision tree descent path refines information about the board position surrounding a candidate move. The results of this work showed that a feature instance with decision tree features is a feasible alternative to the state-of-the-art use of tactical and pattern features in computer Go, in terms of move prediction and playing strength, even though computer Go is a relatively well-developed research area. A move prediction rate of 35.9% was achieved with tactical and decision tree features, and they showed comparable performance to the state of the art when integrated into an MCTS engine with progressive widening. We conclude that the decision tree feature approach shows potential as a method for automatically extracting domain knowledge in new domains. These features can be used to evaluate state-action pairs for guiding searchbased techniques, such as MCTS, or for action-prediction tasks. / AFRIKAANSE OPSOMMING: In rekenaar Go, word skuiwe gewoonlik geselekteer met behulp van ’n boomsoektogalgoritme. Monte-Carlo boomsoektog (MCTS) is tans die dominante algoritme in rekenaar Go. Dit is bekend dat die insluiting van gebiedskennis in MCTS in staat is om die krag van MCTS enjins aansienlik te verbeter. ’n Suksesvolle benadering tot die voorstelling van gebiedskennis in rekenaar Go is taktiek- en patroonkenmerke met geskikte gewigte. Hierdie behels ’n aantal handgemaakte heuristieke en ’n versameling van patrone onderskeidelik. Omdat taktiekkenmerke spesifiek vir Go met die hand gemaak is, en dat patroonkenmerke Go-spesifiek is, is dit nie duidelik hoe hulle maklik oorgedra kan word na ander velde toe nie. Hierdie werk stel dus ’n nuwe verteenwoordiging van gebiedskennis voor, naamlik besluitboomkenmerke. Hierdie kenmerke evalueer ’n toestand-aksie paar deur rekursief die toevoerruimte van toestand-aksie pare te verdeel deur middel van die keuses in die besluitboom, en dan die gewig terug te keer wat ooreenstem met die verdelingselement wat die ooreenstemmende blaarnodus verteenwoordig. In hierdie werk, is besluitboomkenmerke geëvalueer op rekenaar Go, om hul lewensvatbaarheid in vergelyking met veldleidende gebruik van taktiek- en patroonkenmerke te bepaal. In hierdie toepassing van besluitboomkenmerke, verfyn elke navraag in die pad na onder van die besluitboom inligting oor die posisie rondom ’n kandidaatskuif. Die resultate van hierdie werk het getoon dat ’n kenmerkentiteit met besluitboomkenmerke ’n haalbare alternatief is vir die veldleidende gebruik van taktiek- en patroonkenmerke in rekenaar Go in terme van skuifvoorspelling as ook speelkrag, ondanks die feit dat rekenaar Go ’n relatief goedontwikkelde navorsingsgebied is. ’n Skuifvoorspellingskoers van 35.9% is behaal met taktiek- en besluitboomkenmerke, en hulle het vergelykbaar met veldleidende tegnieke presteer wanneer hulle in ’n MCTS enjin met progressiewe uitbreiding geïntegreer is. Ons lei af dat ons voorgestelde besluitboomkenmerke potensiaal toon as ’n metode vir die outomaties onttrek van gebiedskennis in nuwe velde. Hierdie eienskappe kan gebruik word om toestand-aksie pare te evalueer vir die leiding van soektog-gebaseerde tegnieke, soos MCTS, of vir aksie-voorspelling.

Thesis (MSc)--University of Stellenbosch, 2002. / ENGLISH ABSTRACT: In this thesis, we present a technique for the automatic detection of image orientation using Support Vector Machines (SVMs). SVMs are able to handle feature spaces of high dimension and automatically choose the most discriminative features for classification. We investigate the use of various kernels, including heavy tailed RBF kernels. We compare the classification performance of SVMs with the performance of multilayer perceptrons and a Bayesian classifier. Our results show that SVMs out perform both of these methods in the classification of individual images. We also implement an application for the classification of film rolls in a photographic workflow environment with 100% classification accuracy. / AFRIKAANSE OPSOMMING: In hierdie tesis, gebruik ons 'n tegniek vir die automatiese klassifisering van beeldoriëntasie deur middel van Support Vector Machines (SVM's). SVM's kan kenmerkruimtes van 'n hoë dimensie hanteer en kan automaties die mees belangrike kenmerke vir klassifikasie kies. Ons vors die gebruik van verskeie kerne, insluitende RBF-kerne, na. Ons vergelyk die klassifiseringsresultate van SVM's met die van multilaagperseptrone en 'n Bayes-klassifiseerder. Ons bewys dat SVM's beter resultate gee as beide van hierdie metodes vir die klassifikasie van individuele beelde. Ons implementeer ook a toepassing vir die klassifisering van rolle film in a fotografiese werkvloei-omgewing met 100% klassifikasie akuraatheid.

Thesis (MSc (Mathematical Sciences)--University of Stellenbosch, 2006. / Formal methods, such as model checking, have the potential to improve the reliablility of software. Abstract models of systems are subjected to formal analysis, often showing subtle defects not discovered by traditional testing.

Thesis (MSc (Mathematical Sciences. Computer SCience))--University of Stellenbosch, 2009. / In this thesis we consider the dynamic path planning problem for robotics. The dynamic path planning problem, in short, is the task of determining an optimal path, in terms of minimising a given cost function, from one location to another within a known environment of moving obstacles. Our goal is to investigate a number of well-known path planning algorithms, to determine for which circumstances a particular algorithm is best suited, and to propose changes to existing algorithms to make them perform better in dynamic environments. At this stage no thorough comparison of theoretical and actual running times of path planning algorithms exist. Our main goal is to address this shortcoming by comparing some of the wellknown path planning algorithms and our own improvements to these path planning algorithms in a simulation environment. We show that the visibility graph representation of the environment combined with the A* algorithm provides very good results for both path length and computational cost, for a relatively small number of obstacles. As for a grid representation of the environment, we show that the A* algorithm produces good paths in terms of length and the amount of rotation and it requires less computation than dynamic algorithms such as D* and D* Lite.

Thesis (MSc (Computer Science))--University of Stellenbosch, 2009. / Super cially realistic worlds can easily be created using fractal landscapes. Such landscapes are limited in that real-world geological features such as caves and overhangs are impossible using a standard height eld representation. Interactive, realistic simulation of terrain evolution on such landscapes has proved elusive and does not generally provide direct control over the placement of landscape features. This thesis presents an interactive terrain design system, which enables the user to generate localised, speci c terrain features, as well as generalised global characteristics. A fully volumetric terrain representation is used to allow varied three dimensional terrain features. A hybrid rendering algorithm is employed to e ciently render the volumetric data set, while allowing interactive terrain deformation. Small terrain deformation is enabled using voxel carving, while larger terrain features is generated using an adoption of the Wires algorithm. A fractal river network generation technique is presented and adopted for voxel terrain. These tools provide the user with interactive feedback to allow for more dynamic terrain design.