The objective of this thesis research is to model the multiple program objectives used by Defence Research and Development Canada (DRDC) for the annual management and allocation of their broad range of Science and Technology (S&T) projects in order to best achieve the strategic goals of the agency and the government. This M.Sc. thesis presents methodologies, techniques and applications in Linear Programming (LP) and Multi-Criteria Decision Making (MCDM) for decision support in program prioritization and project selection of the DRDC S&T projects. The results of this research produce a model that supports decision makers effectively in the assignment of limited human and financial resources to competing S&T projects based on the evaluation of projects that merit funding and the multiple criteria established by the organization. While there is a well-defined set of criteria for the annual program formulation process, the selection procedure is currently based on simple scoring processes and expert judgement; it lacks a well-defined and structured analysis. The application of an MCDM framework is proposed to take advantage of the well-structured problem and improve annual renewal and ongoing monitoring or project performance measures. The results of the analysis provide a traceable and rigorous MCDM framework to evaluate the performance of DRDC S&T projects for enhanced resource allocation.

Methodical maintenance and renewal of infrastructure systems is critical due to the rapid deterioration of infrastructure assets under increasing loads and environmental effects and the scarcity of resources allocated for their preservation. A crucial step in pavement management is the formation and prioritization of maintenance and rehabilitation (M&R) projects that compete for limited funding for inclusion in the agency’s multiyear pavement management plan (PMPs). In general, many highway agencies perform this task subjectively, and thus a more rational and objective approach is desired to produce sound and justifiable PMPs. Specifically, such methodology should take into account the multiple factors that are considered by engineers in prioritizing M&R projects. This research addresses this need by developing a methodology for use by the Texas Department of Transportation (TxDOT) in preparing their four-year PMPs. Several key decision factors were considered and TxDOT decision makers were surveyed to weigh these factors as to their influence on prioritizing M&R projects. These were then used to develop a priority score for each candidate M&R project. Since TxDOT collects and stores data for individual 0.5-mile pavement sections, these sections must be grouped in a logical scheme to form realistic candidate M&R projects. The incremental benefit-cost analysis was performed on the candidate M&R projects to identify a set of M&R projects that maximizes network’s priority score under budgetary constraint. Future pavement condition was projected using performance prediction models and the process is repeated throughout the planning horizon to produce a multi-year pavement management plan. Data from Bryan district, which consists of 7,075 lane-miles of roadway, were used to develop and validate the PMP methodology. Comparison with the actual PMP (produced by TxDOT) shows some disagreements with the PMP generated by the methodology though the latter was shown to produce more cost-effective and defendable pavement management plans. Since the methodology is founded on TxDOT engineers’ decision criteria and preferences, they can be assured that the PMPs produced by this methodology are in line with their goals and priorities.

The objective of this thesis research is to model the multiple program objectives used by Defence Research and Development Canada (DRDC) for the annual management and allocation of their broad range of Science and Technology (S&T) projects in order to best achieve the strategic goals of the agency and the government. This M.Sc. thesis presents methodologies, techniques and applications in Linear Programming (LP) and Multi-Criteria Decision Making (MCDM) for decision support in program prioritization and project selection of the DRDC S&T projects. The results of this research produce a model that supports decision makers effectively in the assignment of limited human and financial resources to competing S&T projects based on the evaluation of projects that merit funding and the multiple criteria established by the organization. While there is a well-defined set of criteria for the annual program formulation process, the selection procedure is currently based on simple scoring processes and expert judgement; it lacks a well-defined and structured analysis. The application of an MCDM framework is proposed to take advantage of the well-structured problem and improve annual renewal and ongoing monitoring or project performance measures. The results of the analysis provide a traceable and rigorous MCDM framework to evaluate the performance of DRDC S&T projects for enhanced resource allocation.

The objective of this thesis research is to model the multiple program objectives used by Defence Research and Development Canada (DRDC) for the annual management and allocation of their broad range of Science and Technology (S&T) projects in order to best achieve the strategic goals of the agency and the government. This M.Sc. thesis presents methodologies, techniques and applications in Linear Programming (LP) and Multi-Criteria Decision Making (MCDM) for decision support in program prioritization and project selection of the DRDC S&T projects. The results of this research produce a model that supports decision makers effectively in the assignment of limited human and financial resources to competing S&T projects based on the evaluation of projects that merit funding and the multiple criteria established by the organization. While there is a well-defined set of criteria for the annual program formulation process, the selection procedure is currently based on simple scoring processes and expert judgement; it lacks a well-defined and structured analysis. The application of an MCDM framework is proposed to take advantage of the well-structured problem and improve annual renewal and ongoing monitoring or project performance measures. The results of the analysis provide a traceable and rigorous MCDM framework to evaluate the performance of DRDC S&T projects for enhanced resource allocation.

Previous research suggests that people employ priority-related task attributes (e.g., task importance, task length, and task deadline) in prioritization. The process of prioritization employs heuristics to determine task order (Zhang & Feyen, 2007a). These models only address the prioritization process at a task level and do not address the cognitive mechanisms underlying prioritization. Building on previous findings, a process model of prioritization is proposed to explain prioritization during multitasking. Two experiments examined three cognitive processes of prioritization and the influence of time pressure. Three processes were investigated: 1) a process makes magnitude comparisons on priority-related information, 2) a process integrates multiple pieces of information and checks for potential conflicts among information, and 3) a process solves conflicts among priority-related information during prioritization. Under the influence of time pressure, it is hypothesized that people will adopt strategies that require fewer cognitive resources compared to situations where no time pressure exists. A series of task conditions with various configurations of priority-related task attributes was used to illuminate these processes and hypothesis. Hierarchical regression analyses provided evidence for the three cognitive components and suggested that these cognitive components played different roles under time pressure compared to performance under no time pressure. Three fundamental cognitive processes were identified in prioritization and provide implications for personnel selection and training for jobs demanding prioritization and multitasking in the real world.

Recent advances in Next Generation Sequencing (NGS) technologies have facilitated the generation of massive amounts of genomic data which in turn is bringing the promise that personalized medicine will soon become widely available. As a result, there is an increasing pressure to develop computational tools to analyze and interpret genomic data. In this dissertation, we present a systematic approach for interrogating patients’ genomes to identify candidate causal genomic variants of Mendelian and oligogenic diseases. To achieve that, we leverage the use of biomedical data available from extensive biological experiments along with machine learning techniques to build predictive models that rival the currently adopted approaches in the field. We integrate a collection of features representing molecular information about the genomic variants and information derived from biological networks. Furthermore, we incorporate genotype-phenotype relations by exploiting semantic technologies and automated reasoning inferred throughout a cross-species phenotypic ontology network obtained from human, mouse, and zebra fish studies. In our first developed method, named PhenomeNet Variant Predictor (PVP), we perform an extensive evaluation of a large set of synthetic exomes and genomes of diverse Mendelian diseases and phenotypes. Moreover, we evaluate PVP on a set of real patients’ exomes suffering from congenital hypothyroidism. We show that PVP successfully outperforms state-of-the-art methods, and provides a promising tool for accurate variant prioritization for Mendelian diseases. Next, we update the PVP method using a deep neural network architecture as a backbone for learning and illustrate the enhanced performance of the new method, DeepPVP on synthetic exomes and genomes. Furthermore, we propose OligoPVP, an extension of DeepPVP that prioritizes candidate oligogenic combinations in personal exomes and genomes by integrating knowledge from protein-protein interaction networks and we evaluate the performance of OligoPVP on synthetic genomes created by known disease-causing digenic combinations. Finally, we discuss some limitations and future steps for extending the applicability of our proposed methods to identify the genetic underpinning for Mendelian and oligogenic diseases.

No description available.

Feature interactions occur when one feature interferes with the intended operation of another feature. To detect such interactions, each new feature must be tested against existing features. The detected interactions must then be resolved; many existing approaches to resolving interactions require the feature set be prioritized. Unfortunately, the cost to determine a priority ordering for a feature set increases dramatically as the number of features increases. This thesis explores strategies to decrease the cost of prioritizing features, and thus facilitates priority-based solutions to resolving feature interactions. Specifically, this thesis introduces a categorization approach that reduces the complexity of determining priorities for a large set of features by decomposing the prioritization problem. Our categorization approach reduces this cost by using abstraction to divide the system's features into categories based on their main goal or functionality (e.g., block unwanted calls, present call information). Next, in order to detect and resolve the interactions that occur between these seemingly unrelated categories, we identify a set of principles for proper system behaviour that define acceptable behaviour in the global system. For example, a call that should be blocked by a call-screening feature should never result in a voice connection. The categories are then ordered, such that adherence to the principles is optimized. We show that using category priorities, to order a large feature set, correctly resolves interactions between individual features and significantly reduces the cost to determine priority orderings. The four significant contributions that this thesis makes are: 1) the categorization of features, 2) the principles of proper system behaviour, 3) automatic generation of priority orderings for categories, and 4) devising several optimizations that reduce the search space when exploring call simulations during the automatic generation of the priority orderings. These contributions are examined with respect to the telephony domain and result in the identification of 12 feature categories and 9 principles of proper system behaviour. A Prolog model was also created to run call simulations on the categories, using the identified principles as correctness criteria. Our case studies showed the reduced cost of our categorization approach is approximately 1/10^(55) % of the cost of a traditional approach. Given this significant reduction in the cost and the ability of our model to accurately reproduce the manually identified priority orderings, we can confidently argue that our categorization approach was successful. The three main limitations of our categorization approach are: 1) not all features (e.g., 911 features in telephony) can be categorized or some categories will contain a small number of features, 2) the generated priority ordering may still need to be analyzed by a human expert, and 3) the run time for our automatic generation of priority orderings remains factorial with respect to the size of the number of categories. However, these limitations are small in comparison to the savings generated by the categorization approach.

Feature interactions occur when one feature interferes with the intended operation of another feature. To detect such interactions, each new feature must be tested against existing features. The detected interactions must then be resolved; many existing approaches to resolving interactions require the feature set be prioritized. Unfortunately, the cost to determine a priority ordering for a feature set increases dramatically as the number of features increases. This thesis explores strategies to decrease the cost of prioritizing features, and thus facilitates priority-based solutions to resolving feature interactions. Specifically, this thesis introduces a categorization approach that reduces the complexity of determining priorities for a large set of features by decomposing the prioritization problem. Our categorization approach reduces this cost by using abstraction to divide the system's features into categories based on their main goal or functionality (e.g., block unwanted calls, present call information). Next, in order to detect and resolve the interactions that occur between these seemingly unrelated categories, we identify a set of principles for proper system behaviour that define acceptable behaviour in the global system. For example, a call that should be blocked by a call-screening feature should never result in a voice connection. The categories are then ordered, such that adherence to the principles is optimized. We show that using category priorities, to order a large feature set, correctly resolves interactions between individual features and significantly reduces the cost to determine priority orderings. The four significant contributions that this thesis makes are: 1) the categorization of features, 2) the principles of proper system behaviour, 3) automatic generation of priority orderings for categories, and 4) devising several optimizations that reduce the search space when exploring call simulations during the automatic generation of the priority orderings. These contributions are examined with respect to the telephony domain and result in the identification of 12 feature categories and 9 principles of proper system behaviour. A Prolog model was also created to run call simulations on the categories, using the identified principles as correctness criteria. Our case studies showed the reduced cost of our categorization approach is approximately 1/10^(55) % of the cost of a traditional approach. Given this significant reduction in the cost and the ability of our model to accurately reproduce the manually identified priority orderings, we can confidently argue that our categorization approach was successful. The three main limitations of our categorization approach are: 1) not all features (e.g., 911 features in telephony) can be categorized or some categories will contain a small number of features, 2) the generated priority ordering may still need to be analyzed by a human expert, and 3) the run time for our automatic generation of priority orderings remains factorial with respect to the size of the number of categories. However, these limitations are small in comparison to the savings generated by the categorization approach.

Made available in DSpace on 2014-06-12T15:53:11Z (GMT). No. of bitstreams: 2 arquivo1906_1.pdf: 1728491 bytes, checksum: 711dbaf0713ac324ffe904a6dace38d7 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2009 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / Albertins de Lima, Lucas; Cezar Alves Sampaio, Augusto. Test case prioritization based on data reuse for Black-box environments. 2009. Dissertação (Mestrado). Programa de Pós-Graduação em Ciência da Computação, Universidade Federal de Pernambuco, Recife, 2009.