As climate change exacerbates water scarcity and alters available water and fertilizer resources, it is vital that take appropriate measures to ensure sustainable treatment of water, wastewater, and other waste streams that are protective of public health and support recovery and reuse of water and nutrients. The overarching theme of this dissertation is the advancement of next-generation DNA sequencing (NGS) and computational tools for achieving these goals. A suite of relevant fecal and environmental opportunistic pathogens are examined using both culture-based and NGS-based methods. Of particular concern to this research was not only the attenuation and inactivation of pathogens, but also ensuring that optimal treatment approaches reduce antibiotic resistant bacteria (ARB) and antibiotic resistance genes (ARGs). Key systems that were the focus of this effort included nutrient reuse (wastewater-derived biosolids and cattle-derived manure), water reuse, and drinking water systems disrupted by a major hurricane. A field study was carried out to survey a suite of pathogens from source-to tap in six small drinking water systems in Puerto Rico six months after Hurricane Maria. The study revealed that pathogenic Leptospira DNA was detected in all systems that were reliant on surface water. On the other hand, Salmonella spp. was detected in surface and groundwater sources and some distribution system waters both by culture and PCR. The study provided comparison of molecular-, microscopic-, and culture-based analysis for pathogen detection and highlighted the need for disaster preparedness for small water systems, including back-up power supply and access to chlorination as soon as possible after a natural disaster. A second field-study examined wastewater derived solids across an international transect of wastewater treatment plants in order to gain insight into the range of ARG concentrations encountered. It was found that, while total ARGs did not vary between treatment or continent of origin, clinically-relevant ARGs (i.e., ARGs encoding resistance to important classes of antibiotics used in humans) were significantly higher in solids derived from Asian wastewater treatment plants. Estimated loading rates of ARGs to soil under a scenario of land application were determined, highlighting in all cases that they are orders of magnitude higher than in the aqueous effluent. Livestock manure, derived from control cattle and cattle undergoing typical antibiotic treatment, and corresponding composts were also evaluated as common soil amendments in a separate study. In this study, the amendments were applied to two soil types in a greenhouse setting, in order to compare the resulting carriage of ARGs on a root (radish) versus leafy (lettuce) vegetable. Remarkably, radishes were found to harbor the highest relative abundance of total ARGs enumerated by metagenomics, even higher than corresponding soils or manures. Although the total microbial load will be lower on a harvested vegetable, the results suggest that the vegetable surface environment can differentially favor the survival of ARBs. The role of wastewater and water reuse treatment processes in reducing ARB and ARGs was also investigated at field-scale. Two independent wastewater treatment plants both substantially reduced total ARG relative and absolute abundance through biological treatment and settling according to metagenomic analysis. The subsequent water reuse treatment train of one system produced water for non- potable purposes and found further reduction in ARGs after chlorination, but a five hundred percent increase in the relative abundance of ARGs in the subsequent distribution system. In the second plant, which employed a membrane-free ozone-biologically-activated carbon-granular activated carbon treatment train for indirect potable reuse, there were notable increases in total ARG relative abundance following ozonation and chlorination. However, these numbers attenuated below background aquifer levels before recharge. Culture-based analysis of these systems targeting resistant ESKAPE pathogens (Escherichia coil, Staphylococcus aureus, Klebsiella spp., Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterococcus spp.) indicated similar trends as the metagenomic ARG analysis for both systems, but was challenged by sub-optimal media for wastewater samples and low confirmation rates, limiting statistical analysis. In order to advance the application of NGS, molecular, and associated bioinformatic tools for monitoring pathogens and antibiotic resistance in environmental systems, newly emerging methods and field standards for antibiotic resistance assessment were also evaluated. Hybrid assembly, the assembly for both short and long metagenomic sequencing reads, were assessed with an in silico framework in order to determine which available assemblers produced the most accurate and long contigs. Hybrid assembly was found to produce longer and more accurate assemblies at all coverages by reducing error as compared to short read assembly, though the outputs differed in composition from long read assembly. Where it is possible, it is beneficial to sequence using both long- and short-read NGS technologies and employ hybrid assembly, but further validation is recommended. Genome resolved metagenomics has also emerged as a strategy to recover individual bacterial genomes from the mixed metagenomic samples though this is often not well validated. In order to address this, genomes were assembled from reclaimed water systems and were compared against whole-genome sequences of antibiotic resistant E.coli isolates. Metagenome-derived genomes were found to produce similar profiles in wastewater treatment plant influents. A final theme to this dissertation addresses the need to standardize targets, methodologies, and reporting of antibiotic resistance in the environment. A systematic literature review was conducted on assays for the enumeration of key ARGs across aquatic environments and recommendations are summarized for the production of comparable data. In sum, this dissertation advances knowledge about the occurrence of pathogens, ARB, and ARGs across aquatic and agricultural systems and across several countries. Advances are made in the application of NGS tools for environmental monitoring of antibiotic resistance and other targets and a path forward is recommended for continued improvement as both DNA sequencing technologies and computational methodologies continue to rapidly advance. / Doctor of Philosophy / Understanding bacteria in our engineered systems is critical to ensuring drinking water, recycled water, and manure-derived soil amendments are safe for downstream applications. As novel approaches for assessing bacteria are developed, standardized methods and evaluations much be developed to ensure that sound conclusions are made that can appropriately inform policy and practice for the protection of public health. This dissertation focuses on combining bacterial culture and DNA sequencing methods for the study of pathogens (i.e., disease-causing organisms) and antibiotic resistance (i.e., ability of some bacteria to survive antibiotic treatments) in agricultural manure management, water reuse, and drinking water systems. Additionally, this work sought to advance emergent metagenomic analysis tools, which provides a new and potentially powerful pathogen and antibiotic resistance monitoring approach through direct extraction and sequencing of DNA from environmental samples. Antibiotic resistance is a global health challenge and it has been widely recognized that wastewater and agriculture are key control points. When antibiotics are ingested by people or livestock, they select for resistant bacteria in the gut. Mitigation efforts are needed, particularly at wastewater treatment plants and on farms, to ensure that excreted antibiotics and resistant bacteria do not further propagate and pose a risk. However, additional challenges such as climate change have spurred the need for more efficient use of our water and nutrient resources. In this work I examined how nutrient and water reuse treatment methods affect antibiotic resistant bacteria and antibiotic resistance genes using DNA sequencing as well as culture-based methods. In order to assess agricultural practices, a systems approach was conducted at the greenhouse scale to identify key control points to stem the spread of antibiotic resistance when vegetables are grown in soils amended with cattle-derived manure fertilizers. Along the food production chain, vegetables (i.e., radish and lettuce) were found to harbor higher proportions of bacteria carrying antibiotic resistance genes, although the estimated numbers of these bacteria were lower. Solids from an international transect of wastewater treatment plants (Sweden, Switzerland, USA, India, Hong Kong, Phillippenes) were examined because they are also foten used as soil amendments. DNA sequencing of these solids revealed that total measured antibiotic resistance genes did not vary between treatment or continent of origin. Calculations were made to determine the range of total hypothetical outputs of ARGs if the biosolids are land applied. Wastewater reuse systems were also examined using culture and metagenomic DNA analysis so that living pathogens could be compared alongside the total (dead and alive) antibiotic resistance genes. While standard wastewater and subsequent water reuse treatments were found to reduce the absolute numbers of antibiotic resistance genes and bacteria in a treatment plant producing water for non-potable reuse (i.e., irrigation), increases in culturable resistant pathogens and antibiotic resistance genes were apparent in the distribution system (i.e., in the pipes conveying treated water to the point of use). Similar reductions in antibiotic resistant bacteria and resistance genes were also seen in a plant using more advanced treatment (ozonation paired with biofiltration) to produce water suitable for indirect potable reuse via aquifer recharge, but there were indications that ozone and chlorine can increase the proportion of antibiotic resistant bacteria. Finally, genomes were recovered from the metagenomic sequencing analysis and were compared to sequenced culture isolates to validate the capabilities of metagenomic analysis to re-assemble genomes at the strain level, which is often required for pathogen confirmation. Pathogens were also assessed in disrupted drinking water systems in Puerto Rico after Hurricane Maria. Small scale systems that were disrupted by the storm were sampled to identify if pathogens were measurable six months after the hurricane. This work revealed that genes attributed to pathogenic Leptospira were detected in all surface water reliant systems while Salmonella spp. were detected by culture and DNA methods, but only in the source surface and groundwaters, not in the distribution systems delivering water to from the treatment site to the tap. This research also contributed to the advancement of big data analysis pipelines as well as to the standardization of methods to ensure that data produced across studies are comparable. Hybrid assembly, an emergent method that combines both short and long metagenomic DNA sequences generated by different technologies to more accurately recover genomes, was found to improve reliability and accuracy of algorithms aimed at reassembling DNA fragments. Antibiotic resistance is a global challenge, but without standardized methodologies for environmental monitoring, it will be difficult to compare measurements across countries and treatment processes in order to identify effective mitigation strategies. A critical literature review was conducted on assays for the enumeration of key antibiotic resistance genes across aquatic environments so that comparable data can be generated. This will be critical to tap into the tremendous volumes of antibiotic resistance monitoring data being generated around the globe to help identify trends and inform solutions. Collectively, this dissertation advances knowledge about the occurrence of pathogens, antibiotic resistant bacteria and antibiotic resistance genes across aquatic and agricultural systems while also critically evaluating emerging methods for the detection of antibiotic resistance in the environment.

This research develops the availability function for a two component series system in which a component is replaced because of component failure or because it reaches a prescribed age. Also each component replacement provides an opportunity for the replacement of the other component. This last maintenance policy is called an opportunistic replacement strategy. The system functions only if the both components of the system are functioning. The system fails if either of the components fails. Component i is replaced if it fails before attaining age T_i since it was last replaced or maintained. The component i is preventatively maintained if is has not failed by the age T_i. This type of replacement plan is called age replacement policy. When component 1 is being replaced or preventatively maintained, if the age of component j ≠ i exceeds τ_j then both components i and j are replaced at the same time. This type of replacement is called opportunistic replacement of component j and τ_j is called the opportunistic replacement time for component j. The time dependent and long run availability measures for the system are developed. A nested renewal theory approach is used is used to develop the system availability function. The nesting is defined by considering the replacement of a specific one of the components as an elementary renewal event and the simultaneous replacement of both components as the macroscopic renewal event. More specifically, the renewal process for the system represents a starting point for the entire system and is in fact a renewal process. The intervals between system regeneration points are called “major intervals". The age replacement time T_i and opportunistic replacement time τ_i are treated as decision parameters during the model development. The probability distribution of the major interval is developed and the Laplace transform of the system availability is developed. Four replacement models are obtained from the main opportunistic age replacement policy. These are a failure replacement policy, an opportunistic failure model, a partial opportunistic age replacement policy and an opportunistic age replacement policy. These models are obtained as specific cases of the general model. The long run availability measure for the failure replacement model is proven to be the same measure as that developed by Barlow and Proschan. This proof validates the modeling approach. / Ph. D.

The building plumbing microbiome has important implications, especially in terms of its role as a reservoir and conduit for the spread of opportunistic pathogens (OPs), such as Legionella pneumophila. This dissertation applied next-generation DNA sequencing tools to survey the composition of building plumbing microbiomes and assessed hypothetical factors shaping them. A challenge to identifying key factors shaping building plumbing microbiomes is untangling the relative contributions of influent water quality, provided by drinking water utilities, and those of building-level features, such as pipe materials. To this end, standardized pipe rigs were deployed at the treatment plants and in distal portions of the water distribution system at five water utilities across the eastern U.S. Source water and treatment practices appeared to be the overarching factors shaping the microbial taxonomic composition at the tap, with five key water chemistry parameters identified (total chlorine, pH, P, SO42- and Mg2+). Hot water plumbing is of particular interest because OPs tend to proliferate in warm water environments and can be inhaled in aerosols when showering. Two identical lab-scale recirculating hot water rigs were operated in parallel to examine the combined effects of water heater temperature set point, pipe orientation, and water use frequency on the hot water plumbing microbiome. Our results revealed distinct microbial taxonomic compositions between the biofilm and water phases. Importantly, above a threshold of 51 °C, water heater temperature, pipe orientation, and water use frequency together incurred a prominent shift in microbiome composition and L. pneumophila occurrence. While heat shock is a popular means of remediating L. pneumophila contamination in plumbing, its broader effects on the microbiome are unknown. Here, heat shock was applied to acclimated lab-scale hot water rigs. Comparison of pre- versus post- heat shock samples indicated little to no change in either the microbial composition or L. pneumophila levels at the tap, where both water heater temperature and water use frequency had the most dominant effect. Overall, this dissertation contributes to advancing guidance regarding where to most effectively target controls for OPs and also advances research towards identifying the features of a 'healthy' built environment microbiome. / PHD

Annual university rankings produced by mainstream sources, such as U.S. News and World Report, are very popular and viewed as important by a variety of university stakeholders. Consequently, universities expend a great deal of effort in an attempt to ensure they appear in the best possible light. One major component of these ranking systems is the Carnegie Classification of Institutions of Higher Education, which is partly based on the research expenditures reported by the university. This system provides incentives for administrators at institutions of higher education to make strategic accounting choices, with respect to the classification of research expenditures, to improve the prestige of the university. I first measure the amount of accounting discretion within a university's classification of research expenditures and then test whether discretionary research expenditures impact the prestige of a university. Results indicate that discretionary research expenditures are positively associated with university prestige. Specifically, universities within my sample that have positive discretionary research expenditures have an increased probability of subsequently being classified as a Doctoral University with moderate to high research activity by 5% and 7% respectively. In addition, universities within my sample that had positive discretionary research expenditures experienced increases in their ranking of federal funding received relative to other universities by an average of 20.4 positions. These results are consistent with the concept that universities can make certain discretionary accounting choices which can help improve the prestige of the institution with the goal of obtaining additional sources of funding. / Ph. D.

Private wells are understudied potential sources of opportunistic pathogen (OP) infections. OPs, including Legionella and Mycobacterium, are of particular concern for immunocompromised individuals and are known to proliferate in drinking water systems. Much of our knowledge surrounding OP occurrence and growth in drinking water relates to municipal drinking water systems, which primarily use surface water sources and are always treated with primary and secondary disinfection in United States. However, OP occurrence and growth in private wells is not well understood and it is unclear how the knowledge developed in municipal systems will translate to private well systems with rare and infrequent exposure to chemical disinfectants. In addition, because private wells are more susceptible to microbial contamination than municipal systems, the impact of flooding on OP occurrence is of particular concern. Two private well field surveys were conducted to document the incidence of OPs in private well systems. One survey conducted in North Carolina private wells with no history of recent flooding was focused on molecular and culture-based detection of Legionella spp. and Legionella pneumophila. The other survey was a broader molecular (i.e., DNA-based) characterization of the incidence of Legionella spp., L. pneumophila, Mycobacterium spp., Mycobacterium avium (the most commonly nontuberculous mycobacteria associated with disease), and Naegleria fowleri in private wells with recent history of flooding (i.e., Hurricanes Harvey and Irma in 2017, or the Great Louisiana Floods in 2016, extending to Texas, Florida, and Louisiana). All samples in both studies were analyzed for total bacterial 16S rRNA genes, indicator bacteria (e.g., total coliform and Escherichia coli) and inorganic constituents. Information about well system characteristics were obtained through questionnaires sent to participating residents. Widespread detection of OP DNA markers were noted in the flooded well survey. Legionella spp. (detectable in 50-100% of well waters, depending on the flood event) and Mycobacterium spp. (detectable in 13.2-45.0% of well waters) were the most commonly detected among the OPs targets. At the genus level, L. pneumophila (7.9-65.5%) and M. avium (7.9-32.5%) were less commonly detected, but still highly variable. It is not possible to judge whether these OP levels were elevated as a result of the flooding because the sampling was carried out as an emergency response and background levels were not previously tested. Also of interest was whether well characteristics could predict OP levels, including well depth, well type, or treatment. However, none of these emerged as significant predictors of OP detection frequency or levels. Similarly, these OP DNA markers were not elevated in homes reporting submerged wellheads or system damage, suggesting that detection of these OPs is more dependent on the groundwater that supplies these private wells than influx of contaminated surface water. The incidence of DNA markers pertaining to N. fowleri, the "brain eating amoeba" that causes rare incidences of primary amebic meningoencephalitis (PAM), tended to be lower (5.0-12.7%) than that of other OPs targeted, but was more frequently detected in wells reporting submerged well heads, suggesting its occurrence was related to contamination from flood water. A positive association between total bacteria and occurrence of both Legionella spp. and Mycobacterium spp., was observed in private wells of all surveyed areas, contrary to observations in municipal drinking water systems with secondary disinfectant residuals. On the other hand, Legionella reportedly has an optimal growth range of 20-42 °C in municipal systems and recent surveys of municipal systems reported a strong association between Legionella spp. and temperature that was not observed for private wells in this study. We speculate that the essentially "infinite" water age and lack of disinfectant for well water, may contribute to these differences relative to municipal water supplies. The results presented in this work are likely an overestimation of OPs numbers in private wells, as molecular detection of OPs does not distinguish between live and dead cells. In addition, sample sizes were limited by laboratory throughput and budget. Identifying key variables impacting the occurrence of OPs in private wells, given that our study shows that these pathogens are relatively common, might someday help limit the risk of infections. / Master of Science / Non-fecal pathogens that are capable of growth in drinking water systems and causing illness primarily in individuals with compromised immune systems are referred to as opportunistic pathogens (OPs). OPs of concern in drinking water systems in the United States include members of the Legionella genus (i.e., Legionella spp.) and Mycobacterium genus (i.e., Mycobacterium spp.), as well as Naegleria fowleri (also known as “the brain-eating amoeba”). Much of our knowledge surrounding OP occurrence and growth in drinking water relates to municipal drinking water systems. Under the right conditions, OPs have the ability to grow in municipal drinking water systems and in building plumbing systems despite the use of disinfectants. However, OP occurrence and growth in private wells is not well understood and it is unclear how the knowledge developed in municipal systems translates to private well systems that rarely utilize chemical disinfectants. In addition, because private wells are more susceptible to microbial contamination than municipal systems, the impact of flooding on OP occurrence is of particular concern. Two private well surveys were conducted to document the incidence of OP DNA markers and culturable OPs in private well systems. The first survey was conducted in North Carolina private wells with no history of recent flooding and focused on quantification of DNA markers for Legionella spp. and L. pneumophila, as well as culturable L. pneumophila. The second survey was conducted in flood-impacted private wells in vi Texas, Florida, and Louisiana following Hurricanes Harvey and Irma in 2017 and the Louisiana Floods of 2016, quantifying five DNA markers for OPs (Legionella spp., L. pneumophila, Mycobacterium spp., M. avium, and N. fowleri). All water samples in both studies were also analyzed for total bacterial numbers (i.e., total number of copies of a gene present in all bacteria), certain bacteria that indicate environmental and fecal contamination (e.g. total coliform bacteria and Escherichia coli), and inorganics. Information about well system characteristics was obtained through questionnaires sent to participating residents. While it was not possible to sample before the flooding events and determine the background detection rates, this survey found that detection of the target OPs in the flood-impacted wells was widespread, but highly variable. Both culture-based and DNAbased testing methods were used to verify the results because each approach has strengths and weaknesses. Detection of DNA markers indicates the genetic material of the organism being tested for is present and detects DNA of organisms that are both dead and alive. Thus, DNA detection may overestimate the presence live (and infectious) pathogens. Detection of culturable pathogens indicates the organism is alive, but only detects pathogens that readily grow on culture media. There may be other pathogens in the water that do not readily grow on the media. Thus culture may underestimate the occurrence of pathogens. DNA markers for Legionella spp. (detectable in 50-100% of well waters) and Mycobacterium spp. (detectable in 13.2-45.0% of well waters) were the most commonly vii detected among the targets in this study. The detection of DNA markers for L. pneumophila (7.9-65.5%) and M. avium (7.9-32.5%) was less common. There were no private well characteristics, such as well depth, well type, or treatment that emerged as significant predictors of these OP detection or levels. Similarly, these OPs were not elevated in recently flooded homes reporting submerged wellheads or system damage. Thus, detection of these OPs was found to be widespread and sporadic. Detection rates of N. fowleri, which causes rare incidences of primary amebic meningoencephalitis (PAM), DNA was lower (5.0-12.7%) than other OPs, and was also not related to private well characteristics. However, N. fowleri DNA was more frequently detected in wells with wellheads that were submerged due to flooding than in wells with unsubmerged wellheads, as were total coliform bacteria (an indicator of environmental contamination) and total bacterial numbers. This demonstrates concern that submergence of wellheads during flooding is a concern for the introduction of microbial contamination in private wells. This work also explored two trends characteristic of municipal systems that were not observed in these surveys of private wells. First, positive associations between overall bacterial numbers and DNA markers for both Legionella spp. and Mycobacterium spp. were observed in private wells in all surveyed areas. This is contrary to what has been reported for overall bacterial numbers in municipal drinking water systems with a secondary disinfectant residual. Second, Legionella has been known to have an optimal growth range of 32 to 42 °C (90 to 108 °F) in municipal systems and recent surveys of municipal systems with both a free chlorine and chloramine secondary disinfectants viii showed a strong association between Legionella spp. and temperature. These associations were not observed in private wells in this work. Continuous disinfection treatment in municipal drinking water systems may eliminate certain microbes, lowering overall levels of bacteria, while OPs may persist by resisting disinfection, resulting in no correlation between total bacteria and OPs. Private wells do not generally use continuous disinfection and represent low-nutrient environments where naturally occurring OPs can grow alongside other bacteria. The results of this study are likely an overestimation of OPs infection risk associated with private wells, as infection can only be caused by live OPs, which cannot be directly determined by measuring DNA markers. In addition, sample sizes were limited by laboratory throughput and budget. Identifying key variables that impact the occurrence of OPs in private wells is necessary to minimize the risk of associated infections linked to private wells. This work provides strong preliminary evidence that OP occurrence in private wells is relatively commonplace. Science-based options for at-risk (e.g., immunocompromised) individuals to decrease their exposure to OPs in private well water need to be developed.

Diss. (sammanfattning) Stockholm : Karolinska institutet, 2006. / Härtill 5 uppsatser.

In this thesis, the process of optimizing Cooperative Spectrum Sensing in Cognitive Radio has been investigated in fast-fading environments where simulation results have shown that its performance is limited by the Probability of Reporting Errors. By proposing a transmit diversity scheme using Differential space-time block codes (D-STBC) where channel state information (CSI) is not required and regarding multiple pairs of Cognitive Radios (CR's) with single antennas as a virtual MIMO antenna arrays in multiple clusters, Differential space-time coding is applied for the purpose of decision reporting over Rayleigh channels. Both Hard and Soft combination schemes were investigated at the fusion center to reveal performance advantages for Hard combination schemes due to their minimal bandwidth requirements and simplistic implementation. The simulations results show that this optimization process achieves full transmit diversity, albeit with slight performance degradation in terms of power with improvements in performance when compared to conventional Cooperative Spectrum Sensing over non-ideal reporting channels. Further research carried out in this thesis shows performance deficits of Cooperative Spectrum Sensing due to interference on sensing channels of Cognitive Radio. Interference Alignment (IA) being a revolutionary wireless transmission strategy that reduces the impact of interference seems well suited as a strategy that can be used to optimize the performance of Cooperative Spectrum Sensing. The idea of IA is to coordinate multiple transmitters so that their mutual interference aligns at their receivers, facilitating simple interference cancellation techniques. Since its inception, research efforts have primarily been focused on verifying IA's ability to achieve the maximum degrees of freedom (an approximation of sum capacity), developing algorithms for determining alignment solutions and designing transmission strategies that relax the need for perfect alignment but yield better performance. With the increased deployment of wireless services, CR's ability to opportunistically sense and access the unused licensed frequency spectrum, without causing harmful interference to the licensed users becomes increasingly diminished, making the concept of introducing IA in CR a very attractive proposition. For a multiuser multiple-input-multiple-output (MIMO) overlay CR network, a space-time opportunistic IA (ST-OIA) technique has been proposed that allows spectrum sharing between a single primary user (PU) and multiple secondary users (SU) while ensuring zero interference to the PUs. With local CSI available at both the transmitters and receivers of SUs, the PU employs a space-time WF (STWF) algorithm to optimize its transmission and in the process, frees up unused eigenmodes that can be exploited by the SU. STWF achieves higher performance than other WF algorithms at low to moderate signal-to-noise ratio (SNR) regimes, which makes it ideal for implementation in CR networks. The SUs align their transmitted signals in such a way their interference impairs only the PU's unused eigenmodes. For the multiple SUs to further exploit the benefits of Cooperative Spectrum Sensing, it was shown in this thesis that IA would only work when a set of conditions were met. The first condition ensures that the SUs satisfy a zero interference constraint at the PU's receiver by designing their post-processing matrices such that they are orthogonal to the received signal from the PU link. The second condition ensures a zero interference constraint at both the PU and SUs receivers i.e. the constraint ensures that no interference from the SU transmitters is present at the output of the post-processing matrices of its unintended receivers. The third condition caters for the multiple SUs scenario to ensure interference from multiple SUs are aligned along unused eigenmodes. The SU system is assumed to employ a time division multiple access (TDMA) system such that the Principle of Reciprocity is employed towards optimizing the SUs transmission rates. Since aligning multiple SU transmissions at the PU is always limited by availability of spatial dimensions as well as typical user loads, the third condition proposes a user selection algorithm by the fusion centre (FC), where the SUs are grouped into clusters based on their numbers (i.e. two SUs per cluster) and their proximity to the FC, so that they can be aligned at each PU-Rx. This converts the cognitive IA problem into an unconstrained standard IA problem for a general cognitive system. Given the fact that the optimal power allocation algorithms used to optimize the SUs transmission rates turns out to be an optimal beamformer with multiple eigenbeams, this work initially proposes combining the diversity gain property of STBC, the zero-forcing function of IA and beamforming to optimize the SUs transmission rates. However, this solution requires availability of CSI, and to eliminate the need for this, this work then combines the D-STBC scheme with optimal IA precoders (consisting of beamforming and zero-forcing) to maximize the SUs data rates.

Mobile devices are often expected to perform computational tasks that may be beyond their processing or battery capability. Cloud computing techniques have been proposed as a means to offload a mobile device's computation to more powerful resources. In this thesis, we consider the case where powerful computing resources are made available by utilizing vehicles. These vehicles can be repositioned in real time to receive computational tasks from user-carried devices. They can be either equipped with rugged high-performance computers to provide both computation and communication service, or they can be simple message ferries that facilitate communication with a more powerful computing resource. These scenarios find application in challenged environments and may be used in a military or disaster relief settings. It is further enabled by increasing feasibility of (i) constructing a Mobile High Performance Computer (MHPC) using rugged computer hardware with form factors that can be deployed in vehicles and (ii) Message Ferries (MF) that provide communication service in disruption tolerant networks. By analogy to prior work on message ferries and data mules, one can refer to the use of our first schema, MHPCs, as computational ferrying. After illustrating and motivating the computational ferrying concept, we turn our attention into the challenges facing such a deployment. These include the well known challenges of operating an opportunistic and intermittently connected network using message ferries -- such as devising an efficient mobility plan for MHPCs and developing techniques for proximity awareness. In this thesis, first we propose an architecture for the system components to be deployed on the mobile devices and the MHPCs. We then focus on defining and solving the MHPC movement scheduling problem with sufficient generality to describe a number of plausible deployment scenarios. After thorough examination of the MHPC concepts, we propose a scheme in which MHPCs are downgraded to be simple MFs that instead provide communication to a stationary HPC with powerful computing resources. Similar to the MPHCs, we provide a framework for this problem and then describe heuristics to solve it. We conduct a number of experiments that provide an understanding of how the performance of the system using MHPCs or MFs is affected by various parameters. We also provide a thorough comparison of the system in the dimensions of Computation on the Move and Controlling the Mobility.

<p>Electrical production in Sweden today is mainly from nuclear and hydro power. However, there is large increase in renewable energy like wind power and the installed new capacity goals are large. Several electrical production sources are important for the sustainability of the energy system. Maintenance is an approach for keeping a system sustainable. The importance of structured maintenance for reliable electrical production systems triggers the development of qualitative and quantitative maintenance management methods. Examples of these methods are Reliability-Centered Maintenance (RCM) which is a structured qualitative approach that focuses on reliability when planning maintenance, and Reliability Centered Asset Management (RCAM) which is a development of RCM into a quantitative approach with the aim to relate preventive maintenance to total maintenance cost and system reliability.</p><p>This thesis presents models, as applications of RCAM, based on the methods of Life Cycle Cost (LCC) and mathematical optimization, applied to wind and nuclear power plants. Both deterministic and stochastic approaches have been used and the proposed models are based on the <em>Total Cost</em> model, which summarizes costs for maintenance and production loss, and the <em>Aircraft </em>model, which is an opportunistic maintenance optimization model. Opportunistic maintenance is preventive maintenance performed at opportunities. The wind power applications in this study show on different ways to cover costs of condition monitoring systems (CMS) and further on economic benefits of these when uncertainties of times to failure are included in the model. The nuclear power applications show on that the optimization model is dependent on the discount rate and that a high discount rate gives more motivation for opportunistic replacements. When put into a stochastic framework and compared to other maintenance strategies it is shown that an extended opportunistic maintenance optimization model has a good overall performance, and that it, for high values of the constant cost of performing maintenance, is preferable to perform opportunistic maintenance. The proposed models, applied to wind and nuclear power plants, could be extended and adapted to fit other components and systems.</p> / Reliability and cost centered maintenance methods

Les bactéries des microbiotes vivent avec l'homme une interaction mutualiste et tout facteur perturbant l'un ou l'autre des protagonistes peut rompre l'équilibre établi engendrant diverses modifications des interactions existantes. Dans ce contexte instable, les bactéries pathogènes opportunistes d'origine endogène et environnementale qui ont une grande adaptabilité vont pouvoir étendre leur niche écologique ou trouver une nouvelle niche. La diversité des bactéries atypiques ainsi que l'évolution génétique et génomique des bactéries du microbiote respiratoire des patients atteints de mucoviscidose (CF) illustrent ces phénomènes d'adaptation. Le déficit de l'immunité innée locale va en effet être à l'origine d'une colonisation des voies respiratoires (VRCF) par des bactéries endogènes et exogènes qui vont alors mettre en jeu divers mécanismes d'adaptation à cette niche écologique particulière. Le but de notre travail était d'étudier des marqueurs phénotypiques, génétiques et génomiques, témoins potentiels de l'adaptation bactérienne, en particulier au cours de la mucoviscidose. Pour cela, l'étude de la diversité intragénomique des copies du gène de l'ARN ribosomique 16S (rrs), reflétant la capacité d'adaptation bactérienne à des conditions environnementales fluctuantes, a été réalisée par une méthode de PCR et d'électrophorèse avec dénaturation en gradient de température (PCR-TTGE) sur un grand nombre de Veillonella spp. (n=149), d'Achromobacter xylosoxidans (n=164), ainsi que sur 6 autres espèces impliquées dans la colonisation des VRCF (Pseudomonas aeruginosa, Stenotrophomonas maltophilia, Staphylococcus aureus, Haemophilus influenzae, Moraxella catarrhalis et Streptococcus pneumoniae). L'étude de la dynamique du génome de 90 isolats de 12 patients colonisés chroniquement par A. xylosoxidans a été menée par PCR-TTGE, analyse du polymorphisme des fragments de restriction après électrophorèse en champ pulsé et PCR sur séquences répétées. Enfin, l'effet de contraintes environnementales chimiques sur la croissance d'une partie du microbiote CF a été évalué, incluant une étude approfondie de l'effet du NaCl sur 85 isolats de P. aeruginosa.Ces différentes approches nous ont permis de mettre en évidence : i) la diversité intra-génomique des copies de rrs chez Veillonella (74% d'isolats présentant des copies divergentes), H. influenzae (61%), S. maltophilia (38%), A. xylosoxidans (28%), et S. aureus (17%), ii) la clonalité des isolats d'A. xylosoxidans colonisant chroniquement les patients CF associée à une évolution génétique et/ou génomique intra-clonale, iii) l'effet de contraintes environnementales telles que la salinité, le pH et la température sur le microbiote cultivable des VRCF, iv) l'effet antimicrobien du NaCl inhibant la croissance de 100% des P. aeruginosa isolés des VRCF à une concentration de 6%, inférieure à celle utilisée en thérapeutique, et montrant une action bactéricide sur 90% des isolats à une concentration de 10%, v) l'effet multiple du NaCl sur la croissance, le biofilm et la mobilité de P. aeruginosa. Les rrs et le génome constituent des témoins de l'adaptation des bactéries au sein des microbiotes et permettent de mettre en évidence l'importance et la diversité de ces phénomènes dans la niche pathologique que représentent les VRCF. Les contraintes environnementales de la niche écologique influencent cette adaptation. Au cours de la mucoviscidose, le potentiel thérapeutique de la modification de facteurs environnementaux tels que la salinité ou le pH doit être considéré compte tenu de l'impact de ces perturbations sur les communautés microbiennes pathologiques adaptées aux VRCF. / Bacterial microbiotae and human beings developed mutualist interactions. All disrupting factors impacting this equilibrium can modify relationships among microbiota members with diverse consequences. In such an unstable context, opportunistic bacterial pathogens (OBPs) of endogenous or environmental origin showing great adaptability may find favorable conditions leading to ecological niche extension or to new niche colonization. This is illustrated by the diversity of atypical bacteria as well as by genetic and genomic evolution of members of the cystic fibrosis (CF) respiratory tract (CFRT) microbiota. The deficit in local innate immunity allowed colonization by endogenous and exogenous bacteria that will further develop adaptation processes to this particular ecological niche. The aim of this study was to evaluate phenotypic, genetic and genomic potential adaptation markers in different models of OBPs, particularly in CF. For this purpose, we described the variability in the multiple rrs gene copies using PCR and temperature-based denaturing electrophoresis (PCR-TTGE) on large collections of Veillonella (n=149) and Achromobacter xylosoxidans (n=164), as well as on isolates of six other species involved in the CFRT colonization (Pseudomonas aeruginosa, Stenotrophomonas maltophilia, Staphylococcus aureus, Haemophilus influenzae, Moraxella catarrhalis and Streptococcus pneumoniae). Genome dynamics of 90 isolates from 12 patients chronically colonized by A. xylosoxidans was studied by PCR-TTGE, restriction fragment length polymorphism analysis after pulsed-field gel electrophoresis and PCR on repetitive sequences. Finally, the effect of environmental stress was evaluated on part of the growing CF microbiota and a thorough study of NaCl effect on 85 CF P. aeruginosa isolates was performed.These different approaches highlight: i) the intragenomic heterogeneity of the rrs gene copies in Veillonella (74% of isolates), H. influenzae (61%), S. maltophilia (38%), A. xylosoxidans (28%), and S. aureus (17%), ii) a clonal chronic colonization with A. xylosoxidans in 12 CF patients associated with rrs genetic and/or genomic intra-clonal evolution, iii) the effect of environmental stress such as salinity, pH and temperature on the CFRT microbiota, iv) the antimicrobial effect of NaCl on CF P. aeruginosa isolates, a 6% NaCl concentration inhibiting the growth of all the isolates and a bactericidal action being observed for 90% of the isolates with 10% NaCl, v) multiple effects of NaCl on growth, biofilm and mobility of P. aeruginosa. This study shows that rrs genes and genome dynamics witness for bacterial adaptability within microbiota according to environmental constraints and underlines the diversity and importance of adaptation processes in the long-term pathological adapted microbial communities of the CFRT. Modification of environmental factors such as salinity or pH of the CFRT niche may impact the microbiota and should be considered as targets for CF therapeutics.