A software product line (SPL) succeeds because we exploit the similarities between a set of software-intensive systems, together with an understanding of their differences, to reduce development cost, maintenance cost, and user confusion. In SPL engineering, reuse is planned, enabled, and enforced. It is through the development of a set of core assets that reuse is systematically practiced. Requirements assets enhance the effectiveness of reuse since engineers can work on the abstractions closer to the systems' initial concepts. Contemporary SPL requirements engineerin (RE) approaches often adopt the proactive model to build a relatively complete and stable asset base. In practice, the substantial up-front effort and the abrupt transition from existing practices associated with the proactive model present a prohibitive SPL adoption barrier for many organizations that could otherwise benefit. The extractive model overcomes these shortcomings by reusing existing products for the SPL's initial baseline. In this thesis, we present a framework for applying lightweight techniques to extract, model, and analyze a SPL's requirements assets. We define the notion of functional requirements profiles (FRPs) according to the linguistic characterization of a domain's action-oriented concerns, and show that the FRPs can be extracted from a natural language document on the basis of domain-aware lexical affinities that bear a 'verb - direct object' relation. We model the extracted FRPs by analyzing their semantic cases and by extending the orthogonal variability model (OVM). We contribute a set of heuristic rules for uncovering the variation dimensions and dependencies, and discuss merging the OVMs extracted from multiple sources. We relate functional profiles to quality requirements via scenarios, and manage requirements interactions via concept analysis. We present two applications of FRPs to support some other activities in SPL engineering. We conduct several empirical studies to evaluate our framework. The results show that our approach allows the engineers to identify the domain elements more easily and develop the domain models more systematically. Our work fills the void with respect to extracting a SPL's requirements assets, and the main thrust of our work is to promote a set of lightweight, low adoption threshold techniques as a critical enabler for practitioners to capitalize on the order-of-magnitude improvements offered by SPL engineering.

Two experimental seismic surveys were collected in the 1991 LITHOPROBE Trans-Hudson Orogen (THO) data acquisition program. The purpose of the coincident dynamite and vibroseis reflection surveys was to compare crustal images obtained using high-fold low-energy and low-fold high-energy sources. On single-fold field records, signal amplitudes from explosive sources are consistently 50 dB higher than on the corresponding vibroseis records. The vibroseis final stack exhibits better defined upper-crustal reflectivity due primarily to the higher fold. However, at lower-crustal and Moho levels, the dynamite data provides images which are equal or superior to those obtained from the vibroseis data. The dynamite source not only allowed deeper signal penetration but also succeeded in mapping of a number of subcrustal reflections not identified in previous vibroseis data. These new seismic images indicate that the crustal root is not simple depression on the upper-mantle as was inferred initially but a broad (3 s) zone of reflectivity that dips west and extends more than 10 km below the younger regional Moho. Moreover, the dynamite data also indicates that diffraction patterns, detected at lower crustal and Moho depths, have large apertures which permitted proper migration of these lower crustal events. Four vibroseis expanding spread profiles (ESP) were also acquired during the data acquisition program to obtain more detailed and accurate velocity structure. These profiles, with a maximum offset of 18 km, were centered on areas where prominent crustal reflectivity was detected by the regional vibroseis survey. The small source stepout distance (100 m) generated high-fold ($>$30) data. Extensive modeling was carried out to estimate the offset range within which each traveltime approximation and velocity analysis technique may be implemented. The results reveal that velocity estimation becomes more robust and accurate when crustal seismic surveys utilize longer offsets than commonly used. These larger source-receiver separations, however, must be generally limited to offset/depth ratios not exceeding 1.5 when conventional velocity analysis techniques, based on the hyperbolic moveout assumptions, are implemented. Besides the semblance method two velocity estimators adapted to crustal studies, namely the covariance and the $\tau$ - p techniques, were tried. The former yielded the highest resolution followed by the semblance and the $\tau$ - p methods. Resolution of the semblance estimator for a maximum offset of 36 km is equal to that of the covariance method with a corresponding offset of 18 km for mid-crustal reflectors. The advantages provided by the long-offset data acquisition include increased S/N ratio and a greater number of traces with sufficiently large moveouts whichimproved velocity resolution, especially below mid-crustal depths. To achieve similar advantages in a regional crustal reflection survey would require the adoption of longer spread lengths than those presently implemented in standard data acquisition procedures.

The physical redevelopment of Canadian downtown cores has been seen as a primary issue in economically and socially revitalizing urban areas. In the case of Saskatoon, Saskatchewan, the Citys South Downtown area is in need of such rejuvenation. In 2004 redevelopment plans for the area are underway; Saskatoon has set out a proposal to redevelop both its riverfront area and adjacent South Downtown. In order to accomplish the goal of a revitalized South Downtown, the authors of a successful redevelopment proposal must first identify a suitable user population for the area, namely the population of Saskatoon in its entirety, including the Citys disadvantaged central neighbourhood residents. The purpose of this thesis is to define the socio-economic traits of this potential user population for Saskatoons South Downtown in order to recommend facilities and services that should be included in the redevelopment effort. Census data for the fifty-one census tracts that comprise the census metropolitan area of Saskatoon are used to define the social and economic characteristics of this user population. After reviewing the populations socio-economic situations, as well as the social and business organizations that are currently located in the area, recommendations regarding appropriate, requisite facilities and services can be ascertained. These recommendations could then be implemented in the undertakings currently transpiring in Saskatoons South Downtown. Along with defining the socio-economic character of the user population, this study also examines past redevelopment proposals for Saskatoons South Downtown area in an attempt to understand the historical context of the area. The three main past plans for the South Downtown area include: The Meewasin Valley Project (also known as Moriyamas 100 Year Plan), the Mayors Task Force Report, and Princeton Developments South Downtown Master Plan. All of these failed attempts share many common design traits, culminating in the general goal to develop the area into a commercial, residential and recreational area that would cater to the upper-class residents of the city as well as higher-income tourists and visitors to the area. While it must not be assumed that plans which exclude lower-income populations are inherently wrong and destined to be unsuccessful, by targeting such an exclusive population as the primary users of a South Downtown redevelopment, the authors of the previous plans had inadvertently sought to develop an elite district of Saskatoon, financially inaccessible to a vast majority of the citys population. Defining the socio-economic traits of a user population that is comprised of all Saskatonians, and implementing facilities and services that cater to them, would result in an area that is not discriminating; all peoples regardless of life situation or neighbourhood of residence would be able to enjoy an interesting and revitalized South Downtown area of Saskatoon, Saskatchewan.

This dissertation aims at developing and testing a new method that can better capture preferences for multistate health profiles. The motivation arose from the failure of the QALY (Quality-Adjusted Life Year) model in adequately capturing preferences in multistate health profiles. The current QALY-based technique captures preferences for multistate health profiles by evaluating each health state in the profile independently of other states. As the past literature showed, this additive independence condition does not hold in practice and hence such approach is inadequate. To address this issue, this study proposes a novel approach to measure preferences for multistate health profiles by looking at two consecutive health states at a time. It hypothesizes that an evaluation of the future health state is dependent or "conditioned" on the level of the preceding, or current, health state. Characteristics of the current health state that are suspected to impact the resulting conditional preference scores for future health state are systematically explored in a carefully designed empirical study. The interested factors include duration of the current health state, direction of change and amplitude of change between the current and future health states. A 2

The concentrations of seventy-one hydrocarbons (HC) from C2 to C15 were measured simultaneously at two sites in Kaohsiung city in the morning (07-10), the afternoon (13-16), and the evening (18-21) on 14 days in spring 2003. Results show that the most abundant species of Kaohsiung¡¦s air is toluene (43.36-54.49 £gg m-3), followed by i-pentane, 1,2,4-trimethylbenzene, benzene, n-butane, propane and acetylene, in the range 10.36¡V17.11 £gg m-3. The concentrations of 14 halocarbons are in the range 0.25¡V4.57 £gg m-3. Alkanes (around 44.8%) represent the largest proportion of the total HC, followed by aromatics (35.1%), alkenes (15.5%) and halocarbons (5.4%). The afternoon HC concentrations are much lower than those in the morning and at night, due to relatively intense photochemical reaction and favorable dispersion conditions from noon to afternoon. Notable increases in daily HC concentrations are consistent with high temperature, and low HC concentrations on Sunday coincide with low traffic volume. Photochemical activity is investigated, and HC concentrations are found to decline as the NO2/NOx ratio increases. Correlation analyses imply that vehicle exhaust is the dominant source of atmospheric hydrocarbons in Kaohsiung. The profiles of traffic exhausts were also measured for 25 HC species during the morning and afternoon rush hours on four different days in all three traffic tunnels in Kaohsiung City. Results show that VOC concentrations increase with traffic flow rate, and emission profiles in the three tunnels are mostly in the range C2 ¡V C6. Besides the traffic conditions and vehicle type, the pattern of emissions in each tunnel was also influenced by other factors, such as vehicle age, nearby pollution sources, and the spatial or temporal variation of HC in the urban atmosphere. The ozone formation potential (OFP) in each tunnel was assessed based on the maximum incremental reactivities of the organic species, demonstrating that OFP increases with traffic flow rate. Vehicle distribution influences the contributions of organic group to OFP in a tunnel. Meanwhile, when ranked in descending order of contribution to OFP in all tunnels, the organic groups followed the sequence alkenes, aromatics, and alkanes. The possible source categories affecting the atmospheric HC species were further analyzed using factor analysis. Results showed that the major sources of ambient HC at the Nan-Chie and Hsiung-Kong sites are: vehicle exhaust, petrol/diesel exhaust, industrial processes (for example, plastic/rubber process), combustion exhaust, solvent fugitive or business/consume exhaust. Based on the results of factor analysis, source profiles (or fingerprints) were selected and receptor modeling was conducted based on chemical mass balance (CMB). Results of receptor modeling indicated that, at Nan-Chie site, vehicle exhaust (46.33% and 56.36%) represent the largest proportion of total HC, followed by industrial processes (29.63% and 22.37%) in the morning (07-10) and the evening (18-21), respectively; but were industrial process (40.39%) and solvent fugitive exhaust (30.61%) in the afternoon (13-16). Similarly at Hsiung-Kong site, vehicle exhaust (around 46.19% and 49.29%) represent the largest proportion of total HC, followed by industrial processes (23.19% and 26.11%) in the morning and evening, respectively; but were solvent fugitive exhaust (38.85%), vehicle exhaust (28.95%) and industrial process (25.19%) in the afternoon. It is evident that relatively low traffic volumes in the afternoon at both sites reduce the contribution of traffic exhaust to ambient HC.

Today, the healthcare IT vendors develop interfaces for their products to comply with the IHE Profiles and test them with other vendors in annual events, called IHE Connactathons which are one week events organized to perform interoperability testing. Network configuration including hostnames, IP addresses,port numbers for healthcare services and URLs for the various web services used in IHE profiles are stored on online databases accessible from the IHE website. When all companies register for participation to the connectathon, a default configuration is created in the database, assigning default values for healthcare service specific parameters. Afterwards, a company is expected to use the configuration parameters from that database during the connectathon. Except from the connectathons, the vendors could not test their products with other vendors&rsquo / . In this thesis, research on IHE Configuration Management Tool based on IHE profiles, ebXML ebBP, CPP and CPA specifications is presented. The intention behind developing such a tool is to automate the creation and exchange of configuration information between IHE Actors in a Virtual Connectathon Portal. The results of this research enable the IHE product vendors to test their products in interactions with other vendors through the web anytime/anywhere by automatic business service interface configuration. The research enables the healthcare systems to engage with many trading partners at the same time automatically. The work presented in this thesis is realized as a part of IST-027065 RIDE project funded by European Commission.

Protein-protein interactions are important for the prediction of protein functions since two interacting proteins usually have similar functions in a cell. Available protein interaction networks are incomplete / but, they can be used to predict new interactions in a supervised learning framework. However, in the case that the known protein network includes large number of protein pairs, the training time of the machine learning algorithm becomes quite long. In this thesis work, our aim is to predict protein-protein interactions with a known portion of the interaction network. We used Support Vector Machines (SVM) as the machine learning algoritm and used the already known protein pairs in the network. We chose to use phylogenetic profiles of proteins to form the feature vectors required for the learner since the similarity of two proteins in evolution gives a reasonable rating about whether the two proteins interact or not. For large data sets, the training time of SVM becomes quite long, therefore we reduced the data size in a sensible way while we keep approximately the same prediction accuracy. We applied a number of clustering techniques to extract the most representative data and features in a two categorical framework. Knowing that the training data set is a two dimensional matrix, we applied data reduction methods in both dimensions, i.e., both in data size and in feature vector size. We observed that the data clustered by the k-means clustering technique gave superior results in prediction accuracies compared to another data clustering algorithm which was also developed for reducing data size for SVM training. Still the true positive and false positive rates (TPR-FPR) of the training data sets constructed by the two clustering methods did not give satisfying results about which method outperforms the other. On the other hand, we applied feature selection methods on the feature vectors of training data by selecting the most representative features in biological and in statistical meaning. We used phylogenetic tree of organisms to identify the organisms which are evolutionarily significant. Additionally we applied Fisher&sbquo / &Auml / &ocirc / s test method to select the features which are most representative statistically. The accuracy and TPR-FPR values obtained by feature selection methods could not provide to make a certain decision on the performance comparisons. However it can be mentioned that phylogenetic tree method resulted in acceptable prediction values when compared to Fisher&sbquo / &Auml / &ocirc / s test.

This study aims to better understand the characteristics of the tidal- current profiles and the near-bed boundary layer structures off the southwestern coast of Taiwan. The velocity profile is measured by a bottom-mounted ADCP. Six experiments were conducted, each lasted 10~20 days and the water depth ranging 12~18 m. Twenty-minute averaged velocity profiles have been fitted to a logarithmic form with 4% accuracy. The friction velocity (u*) and roughness length (z0) are then derived from the slope and intercept of the best-fitted straight lines. Our results show that the profile shape and friction velocity vary tidally, the latter reaches O(0.06)ms-1 during peak current flow. The magnitude of z0 is large and scattered, but it shows a general trend of decrease with increasing flow speed. The observed log-layer height increases, and the bottom drag coefficient (CD) decrease, respectively with increasing flow speed. Measurements also show that water turbidity increases with rainfall, as a result the z0 and CD also increase. Finally, harmonic analysis of the tidal currents indicate significant changes between winter (homogeneous) and summer (stratified) conditions. In winter the vertical variation of orientation and phase is small, whereas in summer there was a 150 orientation and 250 phase difference (the bottom currents lead the surface currents) between the near surface and near bed regions.

This study determined the polychlorinated dibenzo-p-dioxins/ dibenzo- furans (PCDD/Fs) and 22 metals contents in ash in the super heater (SH), economizer (EC), semi-dryer absorber (SDA), fabric filter (FF), fly ash pit (FAP) and bottom residue (BR) in a municipal solid waste incinerator (KTMSWI). Experimental results showed that average PCDD/FS contents in ash samples from the SH, EC, SDA, FF, FAP and BR were 0.102, 0.788, 0.210, 1.95, 2.04 and 0.0218 ng I-TEQ /g, respectively. PCDD/Fs content was very low in the SH and BR due to high temperatures (around 461¢XC in the SH and 914¢XC in combustion chamber). Conversely, total PCDD/Fs content was significantly high in ash samples from the EC (around 340¢XC), mainly because the temperature is within the favorable range of 250-400¢XC for PCDD/FS formation due to de nova reformation mechanisms. Although the SDA operated at 245¢XC, the PCDD/FS content decreased very significantly, mainly because the temperature was relatively low and because calcium carbonate was introduced into flue gases to dechlorinate and dilute chlorine-containing species. PCDD/Fs were captured by the active carbon in the FF. Furthermore, the duration that fly ash remained in the FF was longer than that for other incinerator units, and thus causing an increasing trend of PCDD/Fs level downstream (except the SDA). Total PCDD/Fs emission factors (£gg /ton-waste; £gg I-TEQ /ton-waste) in ash samples from different units were: SH (42.3; 0.846), EC (326; 6.12 ), SDA (58.1; 1.10), FF (1540; 61.3), FAP (2950; 107) and BR (537; 4.31). Most PCDD/Fs in ash were contributed by the FF (about 56%), and the generation of PCDD/Fs in ash was significant (about 35%) during the transfer process from different units to the FAP. A strong and positive correlation in a logarithmic form existed between PCDD/Fs and chlorine (Cl-) contents in ash. The results showed that principal metals in the incinerator ash were Al, Fe, Zn, Ba, Pb and Cu. On average, these six metals accounted for 96.6%, 96.8%, 97.0%, 94.2% and 96.7% of the total metals in SH, EC, SDA, FF and BR ash, respectively. The emission factors of individual metals from different units were obtained. Volatile metals, such as Cd, Sn, Sb, Hg and Pb, were mostly in fly ash, while lithophilic metals, such as Al, Fe, Ti, V, Cu, Mn, Fe, Co, Ni, Cu, Sr, Mo, Ag, Ba and Cr+6 were mainly in bottom ash. Distribution ratios for total metals in SH, EC, SDA, FF and BR ash were 4.87%, 4.26%, 2.89%, 8.17% and 79.81%, respectively. That is, about 20.2% of total metals were in fly ash and most metals (about 79.8%) were in bottom ash. The Pb content in EC ash and Hg content in FF ash were between alarm contents and maximum legal limits, while Cd and Pb contents in FF ash exceeded the maximum legal limits. The existing transportation system for transferring fly ash from different units to a fly ash pit should be modified for recycling the valuable metals. Ash with less than the alarm content for metals content, such as ash from the SH, SDA and BR, can be reused. Ash with metals exceeding the alarm contents, such as ash from the EC and FF, should be collected and transferred to a different fly ash pit for further treatment, such as metal recycling.

In this thesis, we first investigate the use of automated variance estimators in distribution-free statistical process control (SPC) charts for univariate autocorrelated processes. We introduce two variance estimators---the standardized time series overlapping area estimator and the so-called quick-and-dirty autoregressive estimator---that can be obtained from a training data set and used effectively with distribution-free SPC charts when those charts are applied to processes exhibiting nonnormal responses or correlation between successive responses. In particular, we incorporate the two estimators into DFTC-VE, a new distribution-free tabular CUSUM chart developed for autocorrelated processes; and we compare its performance with other state-of-the-art distribution-free SPC charts. Using either of the two variance estimators, the DFTC-VE outperforms its competitors in terms of both in-control and out-of-control average run lengths when all the competing procedures are tested on the same set of independently sampled realizations of selected autocorrelated processes with normal or nonnormal noise components. Next, we develop WDFTC, a wavelet-based distribution-free CUSUM chart for detecting shifts in the mean of a high-dimensional profile with noisy components that may exhibit nonnormality, variance heterogeneity, or correlation between profile components. A profile describes the relationship between a selected quality characteristic and an input (design) variable over the experimental region. Exploiting a discrete wavelet transform (DWT) of the mean in-control profile, WDFTC selects a reduced-dimension vector of the associated DWT components from which the mean in-control profile can be approximated with minimal weighted relative reconstruction error. Based on randomly sampled Phase I (in-control) profiles, the covariance matrix of the corresponding reduced-dimension DWT vectors is estimated using a matrix-regularization method; then the DWT vectors are aggregated (batched) so that the nonoverlapping batch means of the reduced-dimension DWT vectors have manageable covariances. To monitor shifts in the mean profile during Phase II operation, WDFTC computes a Hotelling's T-square--type statistic from successive nonoverlapping batch means and applies a CUSUM procedure to those statistics, where the associated control limits are evaluated analytically from the Phase I data. We compare WDFTC with other state-of-the-art profile-monitoring charts using both normal and nonnormal noise components having homogeneous or heterogenous variances as well as independent or correlated components; and we show that WDFTC performs well, especially for local shifts of small to medium size, in terms of both in-control and out-of-control average run lengths.