http://localhost:8080/xmlui/handle/123456789/383 Sun, 26 Apr 2026 08:14:36 GMT 2026-04-26T08:14:36Z http://localhost:8080/xmlui/handle/123456789/3488 Title: Design of Resource Management and Energy-Efficient Task Scheduling Algorithms for Fog-Empowered Vehicular Networks Authors: ASIF, THANEDAR MD. Abstract: The delay-sensitive applications, such as self-driving, intelligent transportation, nav igation, and augmented reality assistance, can be evolved in vehicular ad-hoc networks (VANETs) using one of the leading paradigms, fog computing (FC). The demand for these vehicular services has increased due to the emergence of fifth-generation (5G) technology. By blending FC and 5G technologies, the service quality can be improved in intelligent transportation system (ITS) of smart cities. Intelligent vehicles are connected to the road side infrastructure, such as high power nodes (HPNs) and roadside units (RSUs), also called fog nodes (FNs), to obtain on-demand services. These FNs possess finite resources and can provide services to limited vehicles. However, when vehicles reach the network spike in demand, the FNs become impuissant in furnishing services in the existing solutions. As a result, there is a significant reduction in the network service capability and throughput and an increase in the FNs’ energy consumption. Therefore, we propose resource management algorithms such as dynamic resource management (DRM), efficient resource orchestration (ERO) and energy-efficient resource allocation (EERA) to harmonize the resource blocks (RBs) allocation among FNs. Then, to coordinate the allocation of RBs among FNs, the allocated RBs of vehicles in the overlap coverage regions are reduced. This reduction is done by migrating RBs between pairs of FNs to offload upstream services. The problem of reducing RBs among FNs is formulated as integer linear programming (ILP), and its NP-hardness is determined by reducing it from the seminar assignment problem. The proposed algorithm, DRM, considers the set of vehicles in overlapped coverage regions of FNs and communicates with those corresponding FNs. Then, it migrates the RBs of the set of vehicles between pairs of FNs to minimize the allocated RBs. As a result, the network’s service capability is enhanced. The proposed algorithm, ERO, maximizes the network’s throughput by partitioning the FN coverage region into restricted and non restricted coverage regions. Then, it coordinates the allocation of RBs among FNs by reducing RBs for vehicles in the non-restricted coverage regions. A minimum priority queue is constructed using the occupied capacity of FNs to perform optimal migration between pairs of FNs. However, as the vehicles that reach the network grow, FNs’ energy iii consumption increases. Consequently, FNs become futile in delivering services. Therefore, to handle this issue, we present an EERA algorithm to harmonize RB allocation among FNs to reduce the energy utilization of FNs. The proposed algorithm, EERA, relocates the assigned RBs of vehicles in overlap coverage regions amid pairs of FNs, such that the allocated RBs of FNs and energy consumption of FNs are minimized. In ITS, FNs (i.e., HPNs and RSUs) are operated with batteries. FNs are deployed such that the coverage region of each FN intersects with the neighbouring FN(s) to provide services in remote areas where consistent power sources are unavailable. Vehicles in such regions offload delay-sensitive tasks into FNs to get services. However, when the number of vehicles arriving into the network grows over peak hours, the energy dissipation of FNs for processing tasks increases. Consequently, energy-limited FNs become ineffective in delivering services without efficient task scheduling. Therefore, we present reinforcement learning (RL) based energy-efficient and delay-aware (EEDA) task scheduling among FNs in the intersecting regions to reduce the energy dissipation of FNs. The RL agent is trained for different vehicle arrival rates to schedule tasks in a suitable FN of the intersecting areas. The proposed algorithms, DRM, ERO and EERA, are simulated extensively in terms of service capability, serviceability, availability, throughput, energy consumption of FNs and resource utilization. In addition, the simulation results are analogized with benchmark algorithms, such as dynamic resource orchestration (DRO), signal aware (SA), DRO+SA, adaptive resource balance (ARB), minimum cost flow (MCF) and random order (RO), as per their applicability. Similarly, the EEDA algorithm is evaluated by considering FN energy usage, FN response time, and vehicles’ sojourn time in intersecting regions to meet task delay constraints. The simulation outcomes are compared with the priority-aware semi-greedy (PSG), earliest deadline first (EDF), and first come, first serve (FCFS). Description: NITW Mon, 01 Jan 2024 00:00:00 GMT http://localhost:8080/xmlui/handle/123456789/3488 2024-01-01T00:00:00Z http://localhost:8080/xmlui/handle/123456789/3487 Title: Some Metaheuristic Algorithms to Design Deep Neural Networks for Medical Image Detection Authors: Rajesh, Chilukamari Abstract: Medical image analysis plays a critical role in modern healthcare for treatment plan ning, diagnosis, and disease prediction, including brain tumors. It encompasses various modalities, including X-ray, Computed Tomography (CT), and Magnetic Resonance Imag ing (MRI), each with unique strengths and applications. Analyzing these complex im ages pose challenges due to noise, and image quality variability, making them prone to errors and heavily reliant on the knowledge and experience of physicians. Medical imag ing tasks, such as denoising, aim to improve image quality for precise diagnosis, while accurate segmentation enables quantitative analysis and visualization of anatomical abnor malities, with early detection of brain tumors crucial for timely intervention and improved patient outcomes. Deep Neural Networks (DNNs) have shown promising results in medical image analysis. However, manually designing DNN models is challenging, tedious, time consuming, and requires domain-specific knowledge. The increasing number of available training techniques adds complexity to finding the optimal structure, and selecting the suit able hyperparameters for a given task often entails multiple trial-and-error iterations. To address these challenges, Neural Architecture Search (NAS) has emerged as a promising solution, which automates the design of DNNs for specific tasks. Nevertheless, NAS meth ods need further optimization in designing a search space, constructing a DNN from search space (encoding), and exploring different search strategies for specific tasks. The Meta heuristic Algorithm (MA) based methods in NAS have gained traction for automating the DNN architecture design process. This thesis proposes automatically designing flexible and efficient DNN architectures and hyperparameters using MAs. Integrating metaheuris tic optimization with deep learning can lead to adaptable and practical solutions for medical image analysis. Flexible search spaces, advanced techniques, and consideration of compu tational resources contribute to developing practical solutions. The proposed metaheuristic optimization framework has the potential to revolutionize medical image analysis, enhanc ing patient care by enabling better diagnosis, treatment planning, and research in medical imaging. The main objectives of this thesis include: (i) To design a metaheuristic block-based iv deep neural network for medical image denoising, (ii) To develop a metaheuristic-based modified U-shaped network for 2D medical image segmentation with denoising, (iii) To develop a metaheuristic based encoder-decoder model for 3D medical image segmentation, and (iv) To design a multi-objective metaheuristic model for detecting brain tumors in 3D medical images. In this thesis, to achieve the abovementioned objectives, we proposed some metaheuristic-based approaches for medical image analysis tasks, including denoising, seg mentation, and brain tumor detection. Firstly, a metaheuristic block-based deep neural net work is designed for medical image denoising. The denoising performance is enhanced by utilizing the Differential Evolution (DE) algorithm, which facilitates the exploration of various combinations of network components and hyperparameters within the specified search space. Secondly, a metaheuristic-based modified U-shaped network is developed for 2D medical image segmentation with denoising. A modified U-shaped architecture is used with a flexible search space that allows the optimization of individual blocks. Further more, attention blocks are incorporated to enhance segmentation accuracy. The Teaching Learning-Based Optimization (BTLBO) algorithm is used for optimization, resulting in im proved segmentation performance. Thirdly, a metaheuristic-based encoder-decoder model is developed for 3D medical image segmentation. A powerful search space is constructed to optimize the network blocks and training parameters. The Chameleon Search Algorithm (CSA) explores the search space to improve the segmentation performance. Lastly, the third objective is extended to brain tumor detection using a multi-objective optimization ap proach to optimize detection performance and network size. The search space is expanded to include various blocks and parameters. The Multi-objective Iterative Teaching-Learning Based Optimization (MO-ITLBO) algorithm is utilized to identify optimal block structures and training parameters. The experimental results of this research demonstrate the effec tiveness of metaheuristic optimization techniques in enhancing various tasks of medical image analysis. The proposed models outperform existing methods, offering improved de noising, segmentation, and brain tumor detection performance. These advancements can revolutionize medical image analysis, leading to better patient care, diagnosis, and treat ment planning in medical imaging. Description: NITW Mon, 01 Jan 2024 00:00:00 GMT http://localhost:8080/xmlui/handle/123456789/3487 2024-01-01T00:00:00Z http://localhost:8080/xmlui/handle/123456789/3486 Title: Gene Mutations and Motifs Detection for Coronavirus in Biological Sequences of COVID-19 using Deep Learning Models Authors: Gugulothu, Praveen Abstract: In bioinformatics and computational biology, DNA Genome sequence analysis covers a broad range of research issues, such as identifying homology between sequences, recog nition of intrinsic features, mutation detection, genetic diversity disclosure, and species evolution. Sophisticated sequencing technologies produce enormous DNA sequence data, thereby raising the difficulty of analysing sequences as well. The growth of genomic data is much faster compared to the sequence analysis rate. So, there is an enormous need for faster sequence analysis algorithms. Analysis of genome sequences is useful in disease detection, drug development, agriculture and forensics. Our solution to this problem is a Convolutional Neural Network (CNN) that can handle huge DNA sequences using Covid 19 feature extraction. Given the fast spread of the disease, one of the world’s primary concerns is detecting coronavirus disease 2019 (COVID-19). There have been over 1.6 million confirmed in stances of COVID-19, and the disease is rapidly spreading to numerous nations throughout the world, according to recent figures. An analysis of the global incidence and distribution of COVID-19 is presented. We introduce a deep convolutional neural network (CNN) that can distinguish between the original (non-augmented) dataset and the augmented dataset that were both utilised for the assessment. A variety of COVID-19 datasets, including those for MERS-CoV, SARS-CoV, NL63, Alpha-CoV, BetaCoV-1, HKU1-CoV, and 229E-CoV, have been compiled by us from NCBI and GISAD. Each dataset is annotated with its ac cession number and contains nucleotides in FASTA format. In this study, we compiled a positive and negative dataset consisting of 1582 samples with varying genome sequence lengths. By using one-hot encoding, every categorical variable is transformed into its own feature with a binary value of either 1 or 0. Thus, in one-hot encoding, every nucleotide is represented by a four-dimensional one-hot vector; for example, the letters "A," "C," "G," and "T" are encoded as (1, 0, 0, 0), (0, 1, 0, 0), (0, 0, 1, 0)„ and (0, 0, 0, 0), respectively. Us ing the top ten most sick coronavirus sequences as a guide, we trained the suggested CNN module to detect underlying patterns associated with the virus. Learned convolutional fil ters produce motif. The activation values for the 20th filter’s entire sub-sequences are less iii than 0.047075363 and close to 1.086 is the highest activation value is obtained. Advanced Deep Learning Method for COVID-19 Point Mutation Rate Optimisation by Coot-Lion Preventing disease or tailoring treatment to an individual’s needs both de pend on an accurate diagnosis. Unfortunately, the processing time is greatly impacted by the enormous quantity of sequences, even though DNA sequence illness detection is safe. Consequently, computational approaches are suggested to enhance diagnostic precision and expedite the diagnostic procedure. Genetic disorders occur when an organism’s DNA be comes aberrant as a result of mutations in exons. Our new Deep Quantum Neural Network (DQNN) called LBCA-based Deep QNN is built on the Lion-based Coot algorithm. It can forecast the COVID-19 virus using the DNA biological sequence pattern and the rates of point mutations. In this step, the genome sequences undergo feature extraction. This pro cess extracts specific features from the genome sequences, such as CpG-based features and numerical mapping for integer and binary data. Additionally, numerical mapping is applied using the Fourier transform to generate features for skewness, kurtosis, and peak to average power ratio. To get the entropy feature, we also use K-mer extraction. We determined the K-group for point mutations in COVID-19 for both the 200- and 400-genome sequence learning sets, respectively. Afterwards, we also focused on COVID-19 DNA sequence repeats for bi-character and tri-character types, among others, and put forward a DNA sequence clustering model called "ERSIT-GRU" (Exponential Robust Scaling-Identity Tanh-Gated Recurrent Unit) to detect COVID-19 DNA sequence repeats in large datasets. In order to address these challenges, such as the fact that the dataset is tiny, imbalanced, and has fasta quality issues, the dataset has been preprocessed in stages using multiple techniques in order to provide a useful train ing dataset. Consequently, computational approaches are suggested to enhance diagnostic precision and expedite the diagnostic procedure.Genetic disorders manifest in organisms when there is an aberration in their genetic composition as a result of exon mutations. The technique that uses the Trie data structure to forecast disease severity by counting the occur rences of repeat patterns in exons. Due to the tiny database, the suggested method can only forecast the condition of a small number of diseases, despite its effectiveness and speed in doing so based on pattern frequency. There is an immediate need to discover other patterns iv that produce varied diseases in order to solve the problem of a small number of pathogenic patterns. There is data in the genetic code that affects how fast and efficient translation is. In this extensive study of coronaviruses (CoVs) of both human and zoonotic origin, we compare and contrast their codon usage bias, relative errors in insertion and substitution, mutation rates in COVID-19, DNA motif sequence, size, feature extraction based on base frequency, dimer count, and feature extraction based on size. The evolutionary relationship between seven coronaviruses can be shown by the model Harris Hawks Optimisation (HHO) anal ysis, which we have presented. There have been many attempts to fix DNA-based errors using tandem repeats. Depending upon Age, symptoms, and chromosomes all have a role in the different patterns that correlate to normal, pre-mutated, and diseased frequencies. Tandem has identified the ATXN2, DMPK, ATN1, and JPH3 genes, among others, that are involved with disease state. The pattern frequency allows us to predict the disease’s progress and treat it at an early stage. Proposed model reached highest Accuracy in terms of the various Parameters like Accuracy, Precision, Recall, F1 Score. The pattern frequency allows us to predict the disease’s progress and treat it at an early stage. Description: NITW Mon, 01 Jan 2024 00:00:00 GMT http://localhost:8080/xmlui/handle/123456789/3486 2024-01-01T00:00:00Z http://localhost:8080/xmlui/handle/123456789/3485 Title: Design of Early Reliability Prediction Models for Software Projects Authors: Pravali, Manchala Abstract: The significance of software in modern civilization spans across social, political, fi nancial, healthcare, and military domains. However, the increasing complexity and size of software pose challenges, including jeopardizing quality and driving up testing costs. Software reliability is crucial, especially for mission-critical and high-assurance applica tions. This thesis aims to enhance software reliability by predicting faulty modules and estimating development efforts early in the software development lifecycle. Existing pre diction models fall into two categories: Software Reliability Growth Models (SRGMs) and Early Software Reliability Prediction (ESRP) Models. While reliability growth models of fer estimates, relying solely on them for corrective actions can be costly and delayed. Early prediction facilitates refined project planning, timely delivery, and cost overruns, mitigates overestimation and underestimation, allocate resources effectively, and formulates the op timal development approaches. Software fault prediction (SFP), including Within-Project Fault Prediction (WPFP) and Cross-Project Fault Prediction (CPFP) models, improves re liability. Additionally, effort estimation reduces cost estimation uncertainty and enhances software quality by estimating required manpower early in development. The main objectives of this thesis include: (i) To predict the software fault-prone mod ules in within-project through the Weighted Average Centroid based Imbalance Learning approach. (ii) To predict the software fault-prone modules in a cross-project through simi larity based source project and training data selection techniques. (iii) To predict the soft ware fault-prone modules in a cross-project using applicability based source project selec tion, resampling, and feature reduction. (iv) To estimate the software development effort for a project through a two-stage optimization technique. Firstly, a diverse imbalance learning technique is designed for WPFP. The prediction performance is enhanced by diverse synthetic data generation and noisy data elimination. Secondly, this study utilizes the Wilcoxon Signed Rank (WSR) test to identify similar source projects for cross-project prediction. A novel oversampling technique is introduced to address distribution gaps and skewed distributions. Additionally, the performance of CPFP is improved through the use of the Binary-RAO algorithm. This algorithm explores iii diverse combinations of software features and hyperparameters within a specified search space to extract highly correlated features related to module faultiness. Thirdly, the second objective is extended by integrating applicability scores with similarity scores to improve the accuracy of source project selection. A novel resampling technique is devised to extract highly correlated and similar instances while discarding irrelevant ones from the source data, thus enhancing the efficiency of training data construction and reducing distribution discrepancies and class imbalances. Additionally, to tackle the high-dimensionality issue, an efficient deep learning-based Stacked Autoencoder (SAE) model is developed for fea ture reduction, leading to enhanced performance in CPFP. Lastly, an Adaptive Neuro-Fuzzy Inference System (ANFIS) estimation model is developed using multi-objective optimiza tion techniques for efficient software development effort estimation. The multi-objective rank-based improved Social Network Search (SNS) algorithm is applied to extract optimal software project features and to identify ANFIS parameters. The experimental results of this research demonstrate the importance of imbalance learning, source selection, instance selection, and feature optimization for software datasets and the effectiveness of metaheuristic optimization techniques in effort estimation models. The proposed methods outperform existing methods, offering improved software fault pre diction and effort estimation performance, thereby improving reliability prediction overall. Description: NITW Mon, 01 Jan 2024 00:00:00 GMT http://localhost:8080/xmlui/handle/123456789/3485 2024-01-01T00:00:00Z