For modern societies to continue to sustain themselves there needs to be a dissociation between economic growth and environmental degradation or else economic growth will decline consistently together with deteriorating environmental and social health. Various sustainability methodologies can be applied to mitigate against environmental and social degradation. This includes cleaner production which is a proven sustainability methodology that is supported by the United Nations Industrial Development Organisation and the United Nations Environmental Protection Agency in more than eighty countries worldwide. However application of cleaner production practices amongst SMEs has been below expectations where such practices should have become the norm. We have surveyed SMEs in South Africa, where assessments have been carried out by the local National Cleaner Production Centre, to assess SMEs’ perceptions of a range of barrier typologies. Further the barrier methodologies were evaluated to determine whether social responsibility in itself creates a barrier for successful implementation of sustainable practices. This research established that the barrier typologies are more equally balanced than findings in many developed regions. Furthermore, some barriers such as institutional challenges are not as prevalent compared to other developing regions. It was recognized that regulation can be used as an incentive that has an effect on two groupings or axis of barriers identified in this research. Lastly, it was reputed that structured and clear institutional support and strategies further provide enhanced frameworks that were more beneficial than solely focusing on economics for SMEs. / Dissertation (MBA)--University of Pretoria, 2014. / lmgibs2015 / Gordon Institute of Business Science (GIBS) / MBA / Unrestricted

This thesis is submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Ph.D) in the FASTAR group of the Department of Computer Science, University of Pretoria, South Africa. I present a number of algorithms for constructing minimal acyclic deterministic finite automata (MADFAs), most of which I originally derived/designed or co-discovered. Being acyclic, such automata represent finite languages and have proven useful in applications such as spellchecking, virus-searching and text indexing. In many of those applications, the automata grow to billions of states, making them difficult to store without using various compression techniques — the most important of which is minimization. Results from the late 1950’s show that minimization yields a unique automaton (for a given language), and later results show that minimization of acyclic automata is possible in time linear in the number of states. These two results make for a rich area of algorithmics research; automata and algorithmics research are relatively old fields of computing science and the discovery/invention of new algorithms in the field is an exciting result. I present both incremental and nonincremental algorithms. With nonincremental techniques, the unminimized acyclic deterministic finite automaton (ADFA) is first constructed and then minimized. As mentioned above, the unminimized ADFA can be very large indeed — often even too large to fit within the virtual memory space of the computer. As a result, incremental techniques for minimization (i.e. the ADFA is minimized during its construction) become interesting. Incremental algorithms frequently have some overhead: if the unminimized ADFA fits easily within physical memory, it may still be faster to use nonincremental techniques. The presentation used in this thesis has a few unusual characteristics: <ul><li> Few other presentations follow a correctness-by-construction style for presenting and deriving algorithms. The presentations given here include correctness arguments or sketches thereof. </li><li> The presentation is taxonomic — emphasizing the similarities and differences between the algorithms at a fundamental level. </li><li> While it is possible to present these algorithms in a formal-language-theoretic setting, this thesis remains somewhat closer to the actual implementation issues. </li><li> In several chapters, new algorithms and interesting new variants of existing algorithms are presented. </li><li> It gives new presentations of many existing algorithms — all in a common format with common examples. </li><li> There are extensive links to the existing literature. </li></ul> / Thesis (PhD)--University of Pretoria, 2010. / Computer Science / unrestricted

Machine learning algorithms are increasingly used by decision making systems that affect individual lives in a wide variety of ways. Consequently, in recent years concerns have been raised about the social and ethical implications of using such algorithms. Particular concerns include issues surrounding privacy, fairness, and transparency in decision systems. This dissertation introduces new tools and measures for improving the social desirability of data-driven decision systems, and consists of two main parts. The first part provides a useful tool for an important class of decision making algorithms: collaborative filtering in recommender systems. In particular, it introduces the idea of improving socially relevant properties of a recommender system by augmenting the input with additional training data, an approach which is inspired by prior work on data poisoning attacks and adapts them to generate `antidote data' for social good. We provide an algorithmic framework for this strategy and show that it can efficiently improve the polarization and fairness metrics of factorization-based recommender systems. In the second part, we focus on fairness notions that incorporate data inputs used by decision systems. In particular, we draw attention to `data minimization', an existing principle in data protection regulations that restricts a system to use the minimal information that is necessary for performing the task at hand. First, we propose an operationalization for this principle that is based on classification accuracy, and we show how a natural dependence of accuracy on data inputs can be expressed as a trade-off between fair-inputs and fair-outputs. Next, we address the problem of auditing black- box prediction models for data minimization compliance. For this problem, we suggest a metric for data minimization that is based on model instability under simple imputations, and we extend its applicability from a finite sample model to a distributional setting by introducing a probabilistic data minimization guarantee. Finally, assuming limited system queries, we formulate the problem of allocating a query budget to simple imputations for investigating model instability as a multi-armed bandit framework, for which we design efficient exploration strategies.

Multipurpose batch manufacturing systems allow a suite of job types to be processed with a fixed set of machines. These types of systems are commonly found in chemical processing, as well as in computer systems and the service industry. In this thesis we consider the problem of sequencing jobs entering the manufacturing system in order to minimize makespan, or total time to complete processing of the jobs. We formulate this problem as a dynamic programming problem and illustrate the computational difficulty of solving this problem. We give a method for simulation of the system by representing each machine in the system as a finite state automata. This allows one to calculate the makespan given a manufacturing system and a sequence. Due to the complexity of the system, we offer an approximation to the problem. We show that the approximation strategy allows refinement. This progressive refinement of the approximation results in a sequence of approximations that approach the true problem. As the approximation is refined, the computational complexity of the approximated problem grows. For a simplified system, we show that the approximation has bounded error. Furthermore, we show that the error bound of the approximation sequence improves as the approximation approaches the true problem. This presents a trade-off between computational complexity and accuracy of the solution. A decision maker using this sequence of approximations can quickly determine a level of approximation based on the amount of computational power available and the accuracy needed in a solution.

The electric power distribution system usually operates a radial configuration, with tie switches between circuits to provide alternate feeds. The losses would be minimized if all switches were closed, but this is not done because it complicates the system's protection against overcurrents. Whenever a component fails, some of the switches must be operated to restore power to as many customers as possible. As loads vary with time, switch operations may reduce losses in the system. Both of these are applications for reconfiguration. The problem is combinatorial, which precludes algorithms that guarantee a global optimum. Most existing reconfiguration algorithms fall into two categories. In the first, branch exchange, the system operates in a feasible radial configuration and the algorithm opens and closes candidate switches in pairs. In the second, loop cutting, the system is completely meshed and the algorithm opens candidate switches to reach a feasible radial configuration. Reconfiguration algorithms based on linearized transshipment, neural networks, heuristics, genetic algorithms, and simulated annealing have also been reported, but not widely used. These existing reconfiguration algorithms work with a simplified model of the power system, and they handle voltage and current constraints approximately, if at all. The algorithm described here is a constructive method, using a full nonlinear power system model that accurately handles constraints. The system starts with all switches open and all failed components isolated. An optional network power flow provides a lower bound on the losses. Then the algorithm closes one switch at a time to minimize the increase in a merit figure, which is the real loss divided by the apparent load served. The merit figure increases with each switch closing. This principle, called discrete ascent optimal programming (DAOP), has been applied to other power system problems, including economic dispatch and phase balancing. For reconfiguration, the DAOP method's greedy nature is mitigated with a backtracking algorithm. Approximate screening formulas have also been developed for efficient use with partial load flow solutions. This method's main advantage is the accurate treatment of voltage and current constraints, including the effect of control action. One example taken from the literature shows how the DAOP-based algorithm can reach an optimal solution, while adjusting line voltage regulators to satisfy the voltage constraints. / Ph. D.

In the recent years, environmental regulations enforced by federal,state, and local agencies have increasingly addressed water quality issues through progressively more stringent regulations. These regulations have raised concerns in the blue crab industry because processors are now subject to regulations under which new processors are unable to meet the effluent limitations with current processing techniques. This study focuses on the mechanized processing of crab claws. Currently, processors use a brine bath, referred to as the Harris Claw machine, to separate crab claw meat and shell, nd this process yields a waste water which is significantly high in pollutant strength, and is untreatable by biological methods due to the toxicity associated with the high chloride concentrations found in the waste stream. Several alternative crab claw processing systems were developed and evaluated in terms of the meat product yield, the meat product quality, and the wastewater characterization. Two alternatives involved the use of dense media, a 22.5% Staley 1300 corn syrup solution with 5.0% salt and a 30.0% Staley 1300 corn syrup solution, to separate the crab claw meat and shell. These methods, in full scale tests, produced meat yields comparable to that of the brine solution and improved the overall taste of the meat product. However, the effluents had significantly higher BOD5 concentrations. Another alternative to the Harris Claw machine, involved the design, characterization, and testing of a hydraulic separator system (HSS). The HSS was tested on a small scale, but was found to have a meat yield comparable to the Harris Claw machine. The HSS significantly improved the flavor of a final meat product, although the HSS meat product had a significantly lower shelf life than the Harris Claw machine meat product. The waste water quality was improved, because the HSS eliminates the problems associated with a high chloride ion concentration and potentially reduces overall water consumption. / Master of Science

No description available.

No description available.

The project explores alternative strategies to demolition, with the purpose of minimizing waste and optimizing resources in new building projects. The study focuses on the Stockholm suburb of Mariehäll, a former industrial area currently undergoing significant transformation to a mixed-use city. Due to this conversion, wasteful demolitions of large postwar building complexes are prevalent in the area, highlighting the need for sustainable architectural interventions. By examining a set of buildings that are at risk of demolition, the aim of the study was to find a potential host for a new primary school. Among these buildings, a former workshop and warehouse building from the 1950s was identified as the most viable candidate. Through an exploration of reuse and upcycling options, design strategies are proposed for transforming the building into a functional school. By showcasing and analyzing different approaches to the reuse of architectural elements, the study presents methodologies that may be generalized and employed in other projects.

In Bish et al. (2014), two approaches for the generation of hospital evacuation transportation plans were proposed: the minimization of the overall risk and the minimization of the evacuation duration. The resulting evacuation plans differ in terms of overall risk and duration, but also in the evacuation order of patients with different characteristics, the filling of hospital beds, and the assignments of the patients to the various vehicle types. Due to the computational effort of the duration minimization, manipulations of the risk functions for the risk minimization approach were searched in this thesis such that the resulting evacuation plans approach the minimal duration without rules for the assignments of patients to vehicle types. It is possible to create risk functions such that the resulting plans have shorter durations than with the basic risk functions, but the overall risk increases and other properties of the plans change. Furthermore, a new objective function was introduced in this thesis that minimizes an overall penalty function, where penalties are incurred for time intervals in which patients are at the evacuating hospital or being transported. The characteristics of the patients are considered by different weights in the penalty function. For the given problem instance, it is possible to choose penalty factors such that the overall risk is close to the minimal risk or to choose them such that the duration decreases. It is a simple approach with run times that are comparable to the risk minimization approach for the given problem instance. / Master of Science