Institutional-Repository, University of Moratuwa
Welcome to the University of Moratuwa Digital Repository, which houses postgraduate theses and dissertations, research articles presented at conferences by faculties and departments, university-published journal articles and research publications authored by academic staff. This online repository stores, preserves and distributes the University's scholarly work. This service allows University members to share their research with a larger audience.
 Research Publications |  Thesis & Dissertation |  E- Books |
 UoM Journal Publications |  UoM Conference Proceedings |  Articles published in Scimago's Q1 journals |  UoM Research Reports |  Other Articles authored by UoM staff |
Recent Submissions
item: Thesis-Full-text
Energy-efficient cloud task scheduling using deep learning
(2026) Chandrasiri, KDSA; Meedeniya, D
This thesis presents a deep learning-based approach for energy-efficient cloud workflow scheduling by modeling scheduling as a multi-objective reinforcement learning problem that simultaneously optimizes makespan, energy consumption, and Quality of Service (QoS) . The work first introduces GNN-Flat, a baseline Graph Neural Network (GNN) scheduler that represents workflow dependencies using Directed Acyclic Graph (DAG) structures and demonstrates the feasibility of applying graphbased learning to cloud task scheduling. Building on insights from this implementation, the research proposes 2SD-GAT, a two-stage deep reinforcement learning scheduler based on Graph Attention Networks (GAT) , which forms the core contribution of the thesis. The 2SD-GAT architecture separates task selection and resource allocation decisions while leveraging attention mechanisms to capture inter-task dependencies and preference-based reward optimization, enabling improved Pareto trade-offs across competing objectives. Extensive experiments using synthetic and real-world datasets show superior performance compared with heuristic and learning-based baselines. The results show that the proposed approach achieves a 26.8% hypervolume gain with a 4.5x lower Inverted Generational Distance (IGD), along with a makespan improvement of 14.08%. Finally, the research bridges theory and practice by developing a pluggable RL-based scheduling framework, integrations with Apache Airflow and Kubernetes, and supporting tools including workflow engines, execution components, and a web-based GUI, demonstrating how the proposed scheduler can be deployed in realistic cloud orchestration environments.
item: Thesis-Full-text
Development of a system to analyze physiological parameters that are influenced by endothelial dysfunction to assess risks of diabetes
(2026) Samarawickrama, KG; Liyanaarachchi, MR; De Silva , AC; Jayasinghe, S
Vasodilatory Endothelial Dysfunction (VED) is identified as a precursor to cardio- vascular diseases (CVDs) and atherosclerosis. Therefore, early detection of VED is important to prevent the development of severe CVDs and prescribe lifestyle changes or make clinical interventions. Though coronary angiography is the gold standard to detect VED, specialized devices have provided alternative means of detecting VED non-invasively. With these, they have detected VED in patients who are already di- agnosed with CVD or Diabetes Mellitus (DM). Here, we investigated early detection of VED under four subject categories: individuals with pre-diabetes (PDM); healthy individuals with CVD risk factors (HRF); individuals with DM; and individuals with CVD. Non-invasively measured physiological parameters were used to detect VED. We developed a custom-made device to record non-invasive physiological parameters from the distal end of upper limbs namely, Digital Body Temperature, Peripheral Ar- terial Tone, Photo Plethysmography and Peripheral Bio-impedance. Multiple indices which represent unique features of these parameters were calculated considering both long-term and short-term variations. Using these, the performance of this device was validated against the flow-mediated dilation (FMD) procedure in a preliminary study with 10 volunteers (age range: 35 - 60 years): 5 healthy subjects (control) and 5 CVD patients with/without DM. Then the clinical study was conducted with 69 participants (age range: 35-60 years) with all the subject categories: healthy (n=9), HRF (n=15), PDM(n=6),type2DM(n=15),CVDwith/withoutDM(n=24). CVDriskfactorsiden- tified for the HRF category were physical inactivity, unhealthy diet, smoking, alcohol consumption, body mass index, and family history of CVD. Participants with type- 1 DM, liver cirrhosis, renal failure, thyroid disease, spinal cord injuries, and finger deformities were excluded from the study. In the preliminary study, an unpaired t- testperformedbetweenFMDreadingsandindicesderivedfromparameterscalculated from the custom-made device validated device performance. In the clinical study, the ANOVAanalysisshowedpromisingresults(p<0.05)indistinguishinglong-termtrend variations and short-term variations of VED-associated physiological parameters in test subjects diagnosed with CVD or DM along with preclinical stages compared to the healthy conditions. Thereby, we have demonstrated that it may be possible to detect the early development of CVD in preclinical stages through monitoring these VED-associated physiological parameters.
item: Thesis-Abstract
Geotechnical risk assessment of landslides on natural slopes
(2025) Amarasinghe, MP; Kulathilaka, SAS
Landslides are one of the most frequent natural hazards that affect humans, causing significant damage to properties and resulting in fatalities and injuries. Rainfall has been identified as the major triggering factor for landslides, and the tropical region is severely affected due to heavy rainfalls and favourable geological and hydrological conditions prevailing in these regions. Moreover, the landslide impact tends to be severe due to the high economic, political and social vulnerability in the region. However, limited studies have delved into the complex conditions that prevail in these regions. In this thesis, an extensive review was conducted on landslide risk assessment and management. Special emphasis was placed on Sri Lanka as a representative tropical country where landslides are prevalent due to rainfall, complex subsurface conditions, and it is still developing, where many challenges exist, and limited studies have been conducted. This study aimed to assess the landslide susceptibility and hazard for regions with limited data, integrating heuristic methods, process-based methods and machine learning (ML) for landslide prediction, zonation and threshold development. Initially, the applicability of well-established process-based methods was investigated using case studies to predict landslides. Then, using these methods, the critical slopes within a catchment area were identified, and process-based slope stability analysis was conducted, identifying the failure initiation area with extreme event rainfall conditions. Then the runout area for these critical slopes was determined using process-based runout assessment with the initiation area identified. Finally, a framework was developed based on process-based parametric analysis integrated with ML for slope scale prediction of failure and zonation as a quick tool for predicting the landslide hazards. This research delivers robust, scalable, and data-efficient regional landslide risk assessment tools, particularly valuable in data-scarce or resource- constrained settings. Its strong predictive capability and practical adaptability enhance its potential for integration into landslide risk management.
item: Thesis-Abstract
Improving the dyeing affinity of natural dyes for cotton fabrics using nanotechnology : a comparative evaluation of nanoparticles
(2026) Samarawickrama, KGR; Wijayapala, UGS; Fernando, CAN
The growing demand for eco-friendly textile dyeing has renewed interest in natural dyes due to their biodegradability, non-toxicity, and sustainable source. However, their low affinity for cotton fibres and poor colourfastness properties remain major limitations. These challenges are typically addressed using metallic mordants, which increase significant environmental and health concerns. Therefore, the present study aims to enhance the dyeing affinity and fastness properties of natural dyes on cotton fabrics through a nanotechnology-based approach. The objectives of this study are to synthesise and isolate nanoparticles, extraction of natural dyes from selected plant sources, apply nanoparticles as pre-treatments to cotton fabrics, and evaluate their effect on dye performance. Cellulose nanocrystals (CNCs), chitosan nanocrystals (ChsNCs), and iron nanoparticles (FeNPs) were either synthesized or isolated and subsequently characterized using analytical techniques. Natural dye extracts were obtained from the four selected plant leaves of Mangifera indica, Tectona grandis, Lannea coromandelica and Bridelia retusa using aqueous extraction methods, and characterized using analytical techniques. The nanoparticles were used as pre- treatment agents for cotton fabrics prior to dyeing with the natural extracts. The dyed fabrics were assessed for colour properties to evaluate dye uptake and performance. Furthermore, the morphology and interactions among dye molecules, nanoparticles, and cotton fibres were analyzed using scanning electron microscopy (SEM) and Fourier transform infrared (FTIR) spectroscopy. The nanoparticle-treated fabrics showed significantly improved dye performance compared to the untreated sample. Among the nanoparticle treatments, CNCs and ChsNCs exhibited more effective interactions with dye molecules and cotton fibres compared to FeNPs. This study confirms that functionalized nanoparticles offer an efficient, environmentally benign approach to improving the performance of natural dyes on cotton fabrics, thereby enhancing sustainable textile processing and reducing reliance on hazardous mordants and synthetic dyes
item: Thesis-Full-text
Applicability of the adaptive thermal comfort model in naturally ventilated high-rise residential developments in Colombo, Sri Lanka
(2026) Warakapitiya, GYD; Coorey, S; Perera, N
This study investigates the applicability of the ASHRAE 55-2017 adaptive thermal comfort model in naturally ventilated high-rise residential developments in Colombo, Sri Lanka, where rapid urbanization and tropical climatic conditions pose unique challenges for low-income communities. While international standards like ASHRAE 55-2017 provide frameworks for thermal comfort, their relevance in resource- constrained, high-density tropical housing remains underexplored. To address this gap, the research employs a mixed-methods approach, combining longitudinal field measurements of indoor environmental parameters (temperature, humidity, air velocity) with detailed occupant surveys to capture subjective comfort perceptions and adaptive behaviors across five case-study buildings. The results reveal critical discrepancies between model predictions and lived experiences: occupants reported 94.5% dissatisfaction with indoor thermal conditions, despite employing adaptive strategies such as fan use and window adjustments. These differences are aggravated by vertical microclimatic variations, with upper-floor residents experiencing significantly different thermal conditions than lower floors, and socio-cultural constraints, such as privacy concerns that limit effective natural ventilation. Statistical analyses, including Spearman’s rank correlation and ANOVA, further demonstrate that the ASHRAE model underestimates the adaptive capacity of residents in warm- humid climates, particularly during peak discomfort periods (e.g., midday, when 64% of respondents reported severe thermal discomfort). The study concludes that the ASHRAE 55-2017 model requires context-specific recalibration for tropical high- rises. It proposes revised adaptive comfort criteria and emphasizes targeted architectural interventions, such as optimized cross-ventilation and solar shading, alongside policy reforms to address thermal equity in low-income urban housing.