Submitted in fulfillment of the requirements of the Degree of Master of Technology: Chemistry, Durban University of Technology. 2014. / Peptide folding is a very complicated and dynamic process taking place in all living systems. The understanding of a bioactive conformation of the peptides is very important to understand their biological functions and underlying mechanism of action. However, the high flexible nature of peptides makes this process difficult as they can adopt thousands of conformations within the fraction of a second. The usage of experimental techniques in the characterization process is also limited due to several associated complications including synthesis, isolation and crystallization of peptides. The present computational methodologies, on the other hand, are solid enough to provide detailed complementary information about the intrinsic conformational features of peptides by mimicking their physiological conditions. In the present work, molecular dynamics (MD) computational method was used to explore the configurational space of three Aurein peptides, namely Aurein 2.3, Aurein 2.4 and Aurein 2.5. These peptides are secreted by the amphibian skin when they are exposed to external stimuli. These peptides have been reported to possess anti-cancer and anti-bacterial activity with minimum resistance compared to the available drugs. However, despite their medicinal significance, the precise three dimensional structures of Aurein 2.4 and Aurein 2.5 are not as yet known. First, a validation study was performed on Aurein 2.3 to check the efficiency of the computational protocol. The results obtained revealed the presence of -helicity in all residues of the Aurein 2.3, in accordance with its experimental structure. A similar protocol was further used to explore the conformational profiles of the remaining two peptides (Aurein 2.4 and Aurein 2.5) under implicit and explicit solvent conditions. The results obtained revealed that both these peptides exhibit -helical character in all residues although in varying percentages. The -helical region in the case of Aurein 2.4 was localized predominantly in the central residues extending towards its N-terminal residues, whereas it was flanked by N-terminal and the central residues in Aurein 2.5. However, -helicity was completely absent in the explicit solvents, and the peptides preferred to stay either in -turns or extended forms. Hence, the present work provides comprehensive information about the conformational preferences of Aurein peptides which could lead to a better understanding of their native conformations for future investigations and point the way towards developing their new agonists.

There is a growing body of literature reporting the application of hybrid simulations to inform decision making. However, guidance for the design of such models, where the output depends upon more than one modelling paradigm, is limited. The benefits of realising this guidance include facilitating efficiencies in the general modelling process and reduction in project risk (both across measures of time, cost and quality). Focussing on the least well researched modelling combination of agent-based simulation with system dynamics, a combination potentially suited to modelling complex adaptive systems, the research contribution presented here looks to address this shortfall. Within a modelling process, conceptual modelling is linked to model specification via the design transition. Using standards for systems engineering to formally define this transition, a critical review of the published literature reveals that it is frequently documented. However, coverage is inconsistent and consequently it is difficult to draw general conclusions and establish best practice. Therefore, methods for extracting this information, whilst covering a diverse range of application domains, are investigated. A general framework is proposed to consistently represent the content of conceptual models; characterising the key elements of the content and interfaces between them. Integrating this content in an architectural design, design classes are then defined. Building on this analysis, a decision process is introduced that can be used to determine the utility of these design classes. This research is benchmarked against reported design studies considering system dynamics and discrete-event simulation and demonstrated in a case study where each design archetype is implemented. Finally, the potential for future research to extend this guidance to other modelling combinations is discussed.

Specifically equipped and programmed robots are highly successful in controlled industrial environments such as automated production lines. For the transition of robots from such controlled uniform environments to unconstrained household environments with a large range of conditions and variations, a new paradigm is needed to prepare the robots for deployment. Robots need to be able to quickly adapt to their changing environments and learn on their own how to solve their tasks in novel situations. This dissertation focusses on the aspect of learning to predict the success of two-object means-end actions in a developmental way. E.g. the action of bringing one object into reach by pulling another, where the one object is on top of the other. Here it is the “on top” relation that affects the success of the action. Learning the preconditions for complex means-end actions via supervised learning can take several thousand training samples, which is impractical to generate, hence more rapid learning capabilities are necessary. Three contributions of this dissertation are used to solve the learning problem. 1. Inspired by infant psychology this dissertation investigates an approach to intrinsic motivation that is based on active learning, guiding the robot's exploration to create experience useful for improving classification performance. 2. This dissertation introduces histogram based 3D vision features that encode the relative spatial relations between surface points of object pairs, allowing a robot to reliably recognise the important spatial categories that affect means-end action outcomes. 3. Intrinsically encoded experience is extracted into symbolic category knowledge, encoding higher level abstract categories. These symbolic categories are used for knowledge transfer by using them to extend the state space of action precondition learning classifiers. Depending on the actions and their preconditions, the contributions of this dissertation enable a robot to achieve success prediction accuracies above 85% with ten training samples instead of approximately 1000 training samples that would otherwise be required. These results can be achieved when (a) the action preconditions can be easily identified from the used vision features or (b) the action preconditions to be learnt rest upon already existing knowledge, then it is possible to achieve these results by reusing the existing knowledge. This dissertation demonstrates, in simulation, an alternative to handcoding the knowledge required for a robot to interact with and manipulate objects in the environment. It shows that rapid learning, grounded in autonomous exploration, can be feasible if the necessary vision features are constructed and if existing knowledge is consistently reused.

published_or_final_version / Mechanical Engineering / Master / Master of Science in Engineering

published_or_final_version / Environmental Management / Master / Master of Science in Environmental Management

published_or_final_version / Mechanical Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / Education / Master / Master of Science in Information Technology in Education

The contribution of this dissertation is the development of a completely new and accurate algorithm SFSSE for synchronous fault simulation of sequential circuits. The distinctive difference between SFSSE (Synchronous Fault Simulation by Surrogate with Exceptions) and similar approaches for fault simulation in combinational logic circuits is that SFSSE is capable of handling faults stored in more than one memory elements and the reconvergence over time of the stored fault effect with the original fault. The experimental result shows a significant improvement for SFSSE by comparing its execution time to that of parallel fault simulation. After a stored fault list is established during one clock period, all paths from the output of that memory element to the primary outputs might be blocked in subsequent clock periods. A fault is usually propagated through many paths in various subnetworks over several clock periods, and it is detected when only one of these paths reaches a primary output. A new idea for efficiency is suggested in the last chapter to avoid the unproductive simulation activity. In that approach the waste of simulation time is avoided by overlapping the simulation of multiple clock periods.

This study looks into the ability of computer systems, using both video disc images and overlay graphics, to replicate the Army's new M1 tank, so that soldiers using a light pen and simulated test equipment can effectively troubleshoot the tank, using only the technical manual and video terminal display. The computer trained students were then compared with an untrained control group to see if the simulated "hands-on" skills could be effectively transferred to the actual M1 tank during a real troubleshooting exercise. Students' attitudes were also sampled to provide an affective evaluation of the troubleshooting simulation. Research reveals that computer-assisted instruction has had some success in training electrical troubleshooting. The need to train soldiers in mechanical maintenance and the high cost of using the real tank system as a teaching aid led to this study of using computer technology as a cost effective alternative. An analysis of variance showed a significant difference between the training group and the control group in the number of errors made and the number of correct steps per minute at the p > .05 level. An additional test showed that skills learned could be further transferred to a similar, but untrained task on the tank. Students' affective responses indicated that they enjoyed learning from the computer simulation. This technique may have a major implication for civilian education as well. It could be possible to use this interactive video/computer technique to teach 'hands-on' tasks, ranging from chemistry lab experiments to overhauling an automotive transmission. Students could simulate selected tasks, equipment, supplies, and procedures repeatedly to gain mastery in complete safety without consuming valuable resources. Student performance on the computer terminal has been shown to be directly transferable to the actual equipment, and the substitute computer training has been acknowledged by the students as being an enjoyable and worthwhile experience.

No description available.