Multi-Objective Genetic Algorithms for the Single Allocation Hub Location Problem

Hub Location Problems play vital economic roles in transportation and telecommunication networks where goods or people must be efficiently transferred from an origin to a destination point whilst direct origin-destination links are impractical. This work investigates the single allocation hub location problem, and proposes a genetic algorithm (GA) approach for it. The effectiveness of using a single-objective criterion measure for the problem is first explored. Next, a multi-objective GA employing various fitness evaluation strategies such as Pareto ranking, sum of ranks, and weighted sum strategies is presented. The effectiveness of the multi-objective GA is shown by comparison with an Integer Programming strategy, the only other multi-objective approach found in the literature for this problem. Lastly, two new crossover operators are proposed and an empirical study is done using small to large problem instances of the Civil Aeronautics Board (CAB) and Australian Post (AP) data sets.

Preliminary epidemiological study of latent tuberculosis in Mexican agricultural workers in the Niagara region, Canada

It is well documented that the majority of Tuberculosis (TB) cases diagnosed in Canada are related to foreign-bom persons from TB high-burden countries. The Canadian seasonal agricultural workers program (SAWP) operating with Mexico allows migrant workers to enter the country with a temporary work permit for up to 8 months. Preiimnigration screening of these workers by both clinical examination and chest X-ray (CXR) reduces the risk of introducing cases of active pulmonary TB to Canada, but screening for latent TB (LTBI) is not routinely done. Studies carried out in industrialized nations with high immigration from TBendemic countries provide data of lifetime LTBI reactivation of around 10% but little is known about reactivation rates within TB-endemic countries where new infections (or reinfections) may be impossible to distinguish from reactivation. Migrant populations like the SAWP workers who spend considerable amounts of time in both Canada and TBendemic rural areas in Mexico are a unique population in terms of TB epidemiology. However, to our knowledge no studies have been undertaken to explore either the existence of LTBI among Mexican workers, the probability of reactivation or the workers' exposure to TB cases while back in their communities before returning the following season. Being aware of their LTBI status may help workers to exercise healthy behaviours to avoid TB reactivation and therefore continue to access the SAWP. In order to assess the prevalence of LTBI and associated risk factors among Mexican migrant workers a preliminary cross sectional study was designed to involve a convenience sample of the Niagara Region's Mexican workers in 2007. Research ethics clearance was granted by Brock University. Individual questionnaires were administered to collect socio-demographic and TB-related epidemiological data as well as TB knowledge and awareness levels. Cellular immunity to M tuberculosis was assessed by both an Interferon-y release assay (lGRA), QuantiFERON -TB Gold In-Tube (QFf™) and by the tuberculin skin test (TSn using Mantoux. A total of 82 Mexican workers (out of 125 invited) completed the study. Most participants were male (80%) and their age ranged from 22 to 65 years (mean 38.5). The prevalence of LTBI was 34% using TST and 18% using QFTTM. As previously reported, TST (using ~lOmm cut-off) showed a sensitivity of 93.3% and a specificity of 79.1 %. These findings at the moment cannot predict the probability of progression to active TB; only longitudinal cohort studies of this population can ascertain this outcome. However, based on recent publications, lORA positive individuals may have up to 14% probability of reactivation within the next two years. Although according to the SA WP guidelines, all workers undergo TB screening before entering or re-entering Canada, CXR examination requirements showed to be inconsistent for this population: whereas 100% of the workers coming to Canada for the first time reported having the procedure done, only 31 % of returning participants reported having had a CXR in the past year. None of the participants reported ever having a CXR compatible with TB which was consistent with the fact that none had ever been diagnosed with active pulmonary TB and with only 3.6% reporting close contact with a person with active TB in their lifetime. Although Mexico reports that 99% of popUlation is fully immunized against TB within the first year of age, only 85.3% of participants reported receiving BOC vaccine in childhood. Conversely, even when TST is not part of the routine TB screening in endemic countries, a suqDrisingly high 25.6% reported receiving a TST in the past. In regards to TB knowledge and awareness, 74% of the studied population had previous knowledge about (active) TB, 42% correctly identified active TB symptomatology, 4.8% identified the correct route of transmission, 4.8% knew about the existence of LTBI, 3.6% knew that latent TB could reactivate and 48% recognized TB as treatable and curable. Of all variables explored as potential risk factors for LTBI, age was the only one which showed statistical significance. Significant associations could not be proven for other known variables (such as sex, TB contact, history of TB) probably because of the small sample size and the homogeneity of the sample. Screening for LTBI by TST (high sensitivity) followed by confirmation with QFT''"'^ (high specificity) suggests to be a good strategy especially for immigrants from TB high-burden countries. After educational sessions, workers positive for LTBI gained greater knowledge about the signs and symptoms of TB reactivation as well as the risk factors commonly associated with reactivation. Additionally, they were more likely to attend their annual health check up and request a CXR exam to monitor for TB reactivation.

Reading Rocks Junior: A Mixed-Methodological Study of the Efficacy and Accessibility of an Emergent Literacy Intervention

This thesis explored early literacy development in young vulnerable readers. More specifically, this thesis examined an emergent literacy program called Reading Rocks Junior offered by the Learning Disabilities Association of Niagara Region to children four- to six-years of age living in low socioeconomic status communities. Three methodologies were combined to create a rich and complete picture of an effective and accessible literacy program. First of all, a description of the Reading Rocks Junior program is outlined. Secondly, quantitative data that was collected pre- and post- program was analyzed to demonstrate achievement gains made as a result of participating in the program. Finally, qualitative interviews with the program coordinator, the convener of the agency that funded Reading Rocks Junior and three parents whose children participated in the program were analyzed to determine the contextual factors that make Reading Rocks Junior a success.

Feature Selection and Classification Using Age Layered Population Structure Genetic Programming

The curse of dimensionality is a major problem in the fields of machine learning, data mining and knowledge discovery. Exhaustive search for the most optimal subset of relevant features from a high dimensional dataset is NP hard. Sub–optimal population based stochastic algorithms such as GP and GA are good choices for searching through large search spaces, and are usually more feasible than exhaustive and deterministic search algorithms. On the other hand, population based stochastic algorithms often suffer from premature convergence on mediocre sub–optimal solutions. The Age Layered Population Structure (ALPS) is a novel metaheuristic for overcoming the problem of premature convergence in evolutionary algorithms, and for improving search in the fitness landscape. The ALPS paradigm uses an age–measure to control breeding and competition between individuals in the population. This thesis uses a modification of the ALPS GP strategy called Feature Selection ALPS (FSALPS) for feature subset selection and classification of varied supervised learning tasks. FSALPS uses a novel frequency count system to rank features in the GP population based on evolved feature frequencies. The ranked features are translated into probabilities, which are used to control evolutionary processes such as terminal–symbol selection for the construction of GP trees/sub-trees. The FSALPS metaheuristic continuously refines the feature subset selection process whiles simultaneously evolving efficient classifiers through a non–converging evolutionary process that favors selection of features with high discrimination of class labels. We investigated and compared the performance of canonical GP, ALPS and FSALPS on high–dimensional benchmark classification datasets, including a hyperspectral image. Using Tukey’s HSD ANOVA test at a 95% confidence interval, ALPS and FSALPS dominated canonical GP in evolving smaller but efficient trees with less bloat expressions. FSALPS significantly outperformed canonical GP and ALPS and some reported feature selection strategies in related literature on dimensionality reduction.

Feature Selection and Classification Using Age Layered Population Structure Genetic Programming

The curse of dimensionality is a major problem in the fields of machine learning, data mining and knowledge discovery. Exhaustive search for the most optimal subset of relevant features from a high dimensional dataset is NP hard. Sub–optimal population based stochastic algorithms such as GP and GA are good choices for searching through large search spaces, and are usually more feasible than exhaustive and determinis- tic search algorithms. On the other hand, population based stochastic algorithms often suffer from premature convergence on mediocre sub–optimal solutions. The Age Layered Population Structure (ALPS) is a novel meta–heuristic for overcoming the problem of premature convergence in evolutionary algorithms, and for improving search in the fitness landscape. The ALPS paradigm uses an age–measure to control breeding and competition between individuals in the population. This thesis uses a modification of the ALPS GP strategy called Feature Selection ALPS (FSALPS) for feature subset selection and classification of varied supervised learning tasks. FSALPS uses a novel frequency count system to rank features in the GP population based on evolved feature frequencies. The ranked features are translated into probabilities, which are used to control evolutionary processes such as terminal–symbol selection for the construction of GP trees/sub-trees. The FSALPS meta–heuristic continuously refines the feature subset selection process whiles simultaneously evolving efficient classifiers through a non–converging evolutionary process that favors selection of features with high discrimination of class labels. We investigated and compared the performance of canonical GP, ALPS and FSALPS on high–dimensional benchmark classification datasets, including a hyperspectral image. Using Tukey’s HSD ANOVA test at a 95% confidence interval, ALPS and FSALPS dominated canonical GP in evolving smaller but efficient trees with less bloat expressions. FSALPS significantly outperformed canonical GP and ALPS and some reported feature selection strategies in related literature on dimensionality reduction.