oURspace
University of Regina Institutional Repository
The mission of the oURspace digital repository is to share and preserve the scholarly, creative, and cultural work produced at the University of Regina.
What are some of the benefits of depositing your works in oURspace?
- Increased access to your scholarly publications.
- Content is indexed and discoverable in Google Scholar.
- Compliance with open access funding requirements.
- Long term preservation of your work.
Please contact ourspace@uregina.ca if you have questions or want more information about oURspace.
Communities in oURspace
Select a community to browse its collections.
Recent Submissions
Implicit and explicit approaches for efficient healthcare scheduling
(Faculty of Graduate Studies and Research, University of Regina, 2024-03) Ben Said, Aymen; Mouhoub, Malek; Butz, Cory; Khan, Shakil M.; Bais, Abdul; Mohamed, Otmane Ait
Combinatorial optimization problems play a major role in tackling applications such
as healthcare, transportation, education, etc. Solving these applications usually involves
satisfying a set of hard constraints while optimizing one or more objectives.
In this context, exact or approximate methods can be used. While exact methods
guarantee the optimality of the solution returned, they often come with an exponential
running time as opposed to approximate methods that trade the solution’s
quality for a better running time cost. In this context, we tackle the Nurse Scheduling
Problem (NSP). The NSP consists in assigning nurses to daily shifts in a given planning
horizon such that workload constraints are satisfied while hospital’s costs and
nurses’ preferences are optimized. To solve the NSP, we propose implicit and explicit
approaches. In the implicit solving approach, we rely on Machine Learning (ML)
methods based on Association Rules Mining (ARM) and classification algorithms using
historical data (past scheduling solutions) to learn and generate new solutions
through the implicitly learned constraints and objectives that may be embedded in
the learned patterns (e.g., Association Rules, trained ML models, etc). To quantify
the quality of using our implicit approach in capturing the embedded constraints and
objectives in historical data, we rely on the Frobenius Norm (FN). The latter is a
quality measure used to compute the average error between the generated solutions
and historical data. To compensate for the uncertainty related to the implicit approach
given that the constraints and objectives may not be concretely visible in the
produced solutions, we propose an alternative explicit approach where we first model
the NSP using the Constraint Satisfaction Problem (CSP) framework. Then we develop
Stochastic Local Search (SLS) methods and a new Branch and Bound (B&B)
algorithm enhanced with constraint propagation techniques and variables/values ordering
heuristics. Considering that our proposed implicit approach may not guarantee
the feasibility and/or optimality of the generated solution since the constraints and
objectives are represented through the learned patterns from the data, we propose a
data-driven approach to passively learn the NSP as a constraint network from historical
data. The learned constraint network, formulated as a CSP, will then be solved
using the methods we listed earlier.
Application of artificial intelligence to variable rate technology in agriculture
(Faculty of Graduate Studies and Research, University of Regina, 2023-12) Asad, Muhammad Hamza; Bais, Abdul; Al-Anbagi, Irfan; Paranjape, Raman; Hamilton, Howard J.; Shirtliffe, Steven
Variable Rate Application (VRA) plays a pivotal role in enhancing agricultural
profitability by optimizing the use of resources and promoting consistent crop growth.
This also helps mitigate the negative environmental impact of farming practices.
However, the implementation of VRA is heavily reliant on data. An effective VRA
prescription involves an agronomist’s in-depth knowledge of the soil and crop conditions
within Homogenized Management Zones (HMZs). Certain soil attributes like
electrical conductivity, elevation, and soil moisture are measured using proximal sensors
installed on farm machinery. However, other soil properties like soil texture
and Soil Organic Matter (SOM) measurements require soil sampling and laboratorybased
testing. Similarly, crop and weed information is gathered via manual scouting.
The collected SOM, soil texture, and crop information based on limited sampling
may not be representative of whole field conditions resulting in low spatio-temporal
resolution of information. Our research seeks to bridge these gaps by proposing costeffective
and scalable solutions that improve spatio-temporal resolution. We suggest
installing RGB sensors on farm machinery to monitor crop and weed growth, categorize
soil texture, and estimate SOM. This high spatio-temporal information gathered
is subsequently processed to investigate if improved HMZs can be identified. We
develop crop and weed-specific semantic segmentation methods to detect, localize
and quantify crops/weeds, yielding a mean Intersection Over Union (mIOU) up to
83%. These semantic segmentation models are customized to handle agricultural
image data, minimizing memory usage and computational costs during training and
inference. Through this adaptation, we observe a 6% performance improvement in
crop and weed semantic segmentation. The efficiency of binary semantic segmentation
models is further enhanced by up to 12% using ensemble learning methods. We
recognize the strong correlation between soil properties and crop/weed densities and
thus use this relationship to our advantage. We train machine learning models to
predict crop and weed densities based on soil properties and satellite data. To accurately
predict SOM and soil texture from RGB images, we employ a hybrid approach
that combines deep learning and conventional image-processing techniques to overcome
the challenges posed by uncontrolled field conditions data. Lastly, we explore
the potential for identifying HMZs based on resultant high-resolution crop and soil
information. Parts of this research are successfully commercialized under the product
name ”SWATCAM”.
Examining the impact of gender-based microagressions and institutional betrayal on women holding multiple marginalized statuses
(Faculty of Graduate Studies and Research, University of Regina, 2023-09) Anstey, Hannah Jaymes; Klest, Bridget; Gordon, Jennifer; Sangster, Sarah; Carter, Claire; Delker, Brianna C.
Microaggression refers to subtle, often indirect, discriminatory behavior committed
against members of marginalized groups. The subtle nature of microaggressions can
often make them seem innocuous, however, they can have a determinantal impact on
those who experience them. Women are a marginalized group who experience
microaggressions on the basis of gender, rooted in sexism. Further, many women hold
multiple marginalized statuses based on their race, sexuality, disability, age, religion,
body size, or socioeconomic status. As such, they may also experience microaggressions
based on these intersectional identities. While microaggressions in themselves have been
shown to be negatively related to mental health, it has been suggested that how that
initial microaggression is responded to by an institution can cause the mental health
impact to be exacerbated through institutional betrayal (i.e., the failure of an institution
to proactively prevent harm, or to be supportive following harm). Further, if institutions
can respond with courage and support, the impact of a microaggression might also be
changed. Despite the potentially dangerous impact of microaggressions, no research, to
my knowledge, has experimentally investigated the impact of a microaggression directly
after it was perpetrated. The study utilized a mixed-methods approach and investigated
the impact of a single gender-based microaggression with institutional betrayal or
support on state affect in women. Further, the impact of holding multiple marginalized
statuses was investigated. The study took place via Zoom, and participants completed an
unobtrusive measure of state mood directly before and after a male confederate
perpetrated a microaggression consistent with sexual objectification. The experimenter
then responded to said microaggression with institutional support or institutional
betrayal. Once participants recompleted the mood measure, they were asked to complete
a number of questionnaires focused on past experiences of gender-based discrimination
and mental health symptoms and they were asked to respond to qualitative questions
regarding their experience of the microaggression. Results showed that women in the
institutional betrayal condition had a decrease in negative mood after experiencing the
microaggression and women in the institutional support condition had an increase in
negative mood after experiencing the microaggression. Furthermore, regardless of
condition all women showed an increase in positive mood. Future directions and
implications are discussed.
Saturated-unsaturated behavior of natural cohesionless soils
(Faculty of Graduate Studies and Research, University of Regina, 2024-03) Akram, Ilyas; Azam, Shahid; Veawab, Amy; Zeng, Fanhua (Bill); Coulson, Ian; Naggar, Hany El
Cohesionless soils exist across the globe, under various geological settings and climatic regimes, in two distinct natural states. Disintegrated sediments, where soil particles are loosely held together through particle interlocking such as freshly deposited sediments, and intact deposits, in which soil particles are primarily held together through interparticle suction bonding. The geotechnical behavior of such soils is governed by field conditions of applied stress and atmospheric conditions. This research focused on developing a clear understanding of the saturated-unsaturated behavior of cohesionless soils by mimicking the two natural soil states for laboratory investigations (flow-through, volume change, and shear strength) and by using the test results for numerical modeling (two-dimensional and transient seepage-thermal and stability analysis). The main contributions of this research are summarized below.
A simple test method was developed, by utilizing a single sensor and a digital camera, to determine the unsaturated hydraulic conductivity over the entire suction range. The soil exhibited a marginal water holding capacity with air entry value of 8 kPa and a residual suction value of 21 kPa. Following a newly developed sigmoidal function, the soil exhibited a low hydraulic conductivity of 10-7 m/s (saturated value) that gradually decreased with increasing suction (desaturation). Likewise, the difference between the fitted unsaturated hydraulic conductivity values based on upper (10-5 m/s) and lower (10-7 m/s) limits of saturated hydraulic conductivity decreased with suction and converged at vapor conductivity (10-14 m/s).
The conventional oedometer test was improved, by adding a controlled water inflow and a digital data recording, to determine collapse and consolidation. Results showed that with the increase in pre-collapse stress from 25 to 600 kPa, unsaturated compression increased from 0.5 to 5.3% in disintegrated soil and remained close to 0.5% in intact soil. The wetting collapse decreased from 1.1 to 0.1% in disintegrated soil and increased from 6 to 9% in intact soils whereas the total collapse increased from 2 to 5.6% (disintegrated) and from 6 to 9% (intact). The transient volume change during wetting collapse followed a curvilinear trend for both soil states.
The conventional direct shear test was used to determine the shear strength parameters of disintegrated and intact soils under saturated and dried conditions. The disintegrated soil exhibited identical behavior under both saturation states with no clear peak at failure. Apparent cohesion was not observed and friction increased from 44.5° (saturated) to 48°(unsaturated). The intact soil behaved similar to the disintegrated soil in saturated state due to the absence of suction and had a clear peak and residual similar to dense soils. Apparent cohesion and friction angle respectively increased from 0 kPa and 42° (saturated) to 91 kPa and 36° (unsaturated).
A transient and two-dimensional seepage-thermal model was developed to determine the stability of typical embankments with a low slope (18 m) and high slope (26 m). These slopes were analyzed under mean, extreme wet, and extreme dry climatic conditions along with four ponding conditions (none, upstream, downstream, both) with and without vehicular loading.
The laboratory protocols and the numerical model are crucial for shallow and young geological deposits that are in direct contact with the atmosphere and where most civil infrastructure resides. The findings of this research are useful for the near design, construction, and rehabilitation of urban facilities exposed to climatic change impacts.
Kiannet: An attention-based CNN-RNN model for violence detection
(Faculty of Graduate Studies and Research, University of Regina, 2024-04) Ahmadi Vosta Kolaei, Soheil; Yow, Kin-Choong; Chan, Christine; Maciag, Timothy; Zilles, Sandra; Eramian, Mark
Violent behaviour poses a significant risk to societal stability and public safety. As
part of proactive strategies to counteract this threat, many organizations and institutions
have implemented surveillance systems to monitor and identify potential violent
instances. Nevertheless, manual review and analysis of vast surveillance footage can
be a daunting and error-prone task for human operators, necessitating the advent of
automated systems for efficient and precise violence detection.
This study introduces a novel approach for violence detection composed of a CNNRNN
structure based on an attention mechanism for binary and multi-class classification
of abnormal behaviours. We called our proposed model KianNet because Kian
is the name of an intelligent innocent murdered in a violent incident, and we chose
his name as a representative of all people who suffered from violent behaviours. In
this technique, a CNN-RNN structure is applied to an input video to extract features
from a sequence of frames and by adding a combination of Multi-Head Self-Attention
(MHSA) and ConvLSTM layers, it can detect the violent event and determine the
type of the observed anomaly.
The key to KianNet’s performance is implementing the MHSA layer, which allows
the model to focus on specific spatiotemporal regions of relevance, improving its capacity
to differentiate between normal and violent events. Consequently, the MHSA
layer boosts KianNet’s discriminatory power, enabling it to discern violent incidents
from regular activities better. Through empirical evaluations, KianNet has proven its
superior performance in violence detection tasks. Our findings reveal that KianNet
outperforms its closest competitors’ accuracy by roughly 10 percent. This substantial
performance margin substantiates the robustness and reliability of KianNet, cementing
its potential as an effective tool in automated surveillance systems for violence
detection.