GWO fuzzy PID controlled cuk and SEPIC converter based PFR Alok Kumar Mishra, Pradip Kumar Nanda, Shekharesh Barik, Dipansu Ranjan Mohapatra, Pratap Kumar Sahoo, Akshaya Kumar Patra Intelligent Computing Techniques and Applications, 2025 This article gives a general presentation investigation of two power factor (PF) rectification (PFR) converter (PFRC) geographies: Cuk type and SEPIC type converter. Again average current type control (ACC) technique is utilized here. Notwithstanding, for control of output voltage, techniques like conventional PID regulator (PIDR) or fuzzy PID regulator (FPIDR) is used and the regulator gains is calculated by Grey Wolf Optimization (GWO) process, considering integral of time biased absolute error (ITAE). The analysis of both PFRC acted in Simulink/MATLAB and the anticipated geographies are studied under consistent state and vibrant state situations.
YOLO & ML based crack detection and strength prediction for structural health monitoring of bridges Surajit Mohanthy, Niva Tripathy, Shekharesh Barik, Rojalin Dash, Rajeev Agrawal, Prabhu Prasad Nanda Intelligent Computing Techniques and Applications, 2025 Machine learning (ML) and You Only Look Once (YOLO)-based approaches have shown significant promise in crack detection and strength prediction for structural health monitoring (SHM) of bridges. These techniques offer improved accuracy, real-time monitoring capabilities, and efficient data processing for bridge maintenance and safety. YOLO-based algorithms have been successfully applied to bridge crack detection, demonstrating high accuracy and real-time perfor-mance.
PQ improvement with TLBO FOPID based SHAF Alok Kumar Mishra, Jeevan Jyoti Mahakud, Shekharesh Barik, Pratap Kumar Sahoo, Dipansu Ranjan Mohapatra, Akshaya K. Patra Intelligent Computing Techniques and Applications, 2025 In this work a TLBO FOPID shunt hybrid active filter (SHAF) designed to compensate harmonics and reactive power is presented. The hybrid approach combines shunt passive and active filters, leveraging the advantages of each component. Passive filters (PF) are avoided due to their massive and multifaceted strategy, while active filters (AF) are costly for high ratings. Simulink prototypes of SHAF are created to achieve fewer misleading sine wave input currents. In this approach, extraction of current reference for the AF diverges from conformist ways for instance (p-q) or (id-iq) theory, eliminating the need for load current sensing. A unique controller, i.e. fractional order PIDC (FOPIDC), is employed to estimate the crest current reference for SHAF, optimizing the parameters of the controller teaching learning-based optimization (TLBO) process is adopted. The algorithm incorporates integration time weighted absolute error (ITAE). Constraints like power factor (PF), THD, P the active power and Q the reactive power are assessed through simulation, and predicted system is validated under transitory and steady conditions for the optimal FOPID and PID controller (OFOPIDC & OPIDC), utilized for current reference extraction.
Optimizing Cloud Load Balancing Performance Using Hybrid Ant Colony and Artificial Bee Colony Optimization Technique Yennam Praneetha, Madishetty Vignesh, Kamsani Akash, Saroja Kumar Rout, V N L N Murthy, Shekharesh Barik 2025 Global Conference on Information Technology and Communication Networks Gitcon 2025, 2025 In the quickly changing world of cloud computing, efficient load balancing is crucial for improving customer happiness and resource efficiency. This article suggests a potent hybrid model that combines the Artificial Bee Colony (ABC) optimization methodology with the Ant Colony Optimization (ACO) method to solve load balancing issues. The technique replicates the ants' and honeybees' natural search patterns by dynamically distributing workload among virtual machines (VMs), which leads to balanced loads, quicker reaction times, and greater resource efficiency. Simulations demonstrate that the Hybrid ACO-ABC algorithm outperforms traditional methods in throughput, latency, and energy efficiency. In this paper, the ability of nature-inspired solutions to optimize cloud infrastructure efficiency is proven by providing a scalable and robust load-balancing system that adjusts to varying workloads and resource availability.
Optimizing Cloud Computational Performance via Hybridizing Rock Hyrax and Grey Wolf Optimization Algorithms Saroja Kumar Rout, Kottu Santosh Kumar, Pradyumna Kumar Mohapatra, Shekharesh Barik, Muhammed Basheer Jasser, Samuel-Soma M. Ajibade, Hui Na Chua 2025 IEEE International Conference on Automatic Control and Intelligent Systems I2cacis 2025 Proceedings, 2025 Task scheduling in cloud computing presents major challenges due to dynamic workloads, resource heterogeneity, and the inherent limitations of traditional optimization methods in maintaining an effective balance between exploration and exploitation. Existing algorithms like Rock Hyrax Optimization (RHO) and Grey Wolf Optimization (GWO) have individual drawbacks: RHO struggles with slower convergence speeds, while GWO often risks premature convergence, resulting in inefficient scheduling outcomes. To overcome these limitations, this research focuses on developing a more adaptive and efficient scheduling model suited for cloud data centers, edge computing, and real-time applications. A novel hybrid RHO-GWO optimization model is proposed, which dynamically balances exploration and exploitation based on the nature of workload conditions. The model intelligently adapts, selecting the optimal strategy during task scheduling to enhance overall performance. Implementation and evaluation were conducted within simulated cloud environments, assessing key performance metrics such as makespan, resource utilization, load balancing, and energy efficiency. The results demonstrate that the Hybrid RHO-GWO algorithm outperforms RHO and GWO across key metrics. It achieves, the lowest makespan (65s–160s), the highest resource efficiency (0.91), the best load balancing (variance: 0.050–0.053 and average variance at 0.88), and the lowest energy consumption (950W), offering up to 20.83% energy savings, making it the most efficient and balanced solution among the three.
A Novel Cyclone Detection System Using Deep Learning Shekharesh Barik, Rishub Kumar, Siba Narayan Sahoo, Biswa Ranjan Jit 2024 IEEE International Conference on Information Technology Electronics and Intelligent Communication Systems Iciteics 2024, 2024 Cyclones as natural disasters, pose significant threats to human life and property, necessitating precise and timely cyclone detection systems. Traditional detection approaches have limitations in accuracy and speed, resulting in potential delays in issuing early warnings. In response to the challenges posed by traditional detection methods, this research introduces an innovative Cyclone Detection System (CDS) that leverages machine learning and deep learning algorithms to significantly enhance cyclone detection efficiency and reliability. The CSD integrates diverse meteorological data sources to construct a comprehensive and multi-dimensional feature set. This feature set aims to capture the intricate and dynamic patterns associated with cyclone formation and development. Advanced machine learning and deep learning techniques, specifically Convolutional Neural Networks and Recurrent Neural Networks, are employed to process the feature set and accurately classify cyclonic patterns. The deep learning models utilized in the CDS facilitate automatic learning and adaptation to evolving cyclone characteristics, greatly improving its detection capabilities. To evaluate the effectiveness of the proposed system, extensive experiments were conducted using historical cyclone datasets. The results affirm the superiority of the CDS over traditional methods, achieving significantly higher accuracy i.e. 92% in cyclone detection. Furthermore, the real-time implementation of the CDS ensures swift detection and monitoring of cyclones, enabling the timely issuance of early warnings to vulnerable regions. By providing precise and timely information, this groundbreaking CDS has the potential to save lives, mitigate property damage, and enhance disaster response coordination.
Enhancing Heart Disease Prediction Accuracy Through Hybrid Machine Learning Methods Nukala Sujata Gupta, Saroja Kumar Rout, Shekharesh Barik, Ruth Ramya Kalangi, B Swampa Eai Endorsed Transactions on Internet of Things, 2024 INTRODUCTION: Over the past few decades, heart disorders have been the leading cause of mortality worldwide. People over 55 must get a thorough cardiovascular examination to prevent heart disease or coronary sickness and identify early warning signs. To increase the ability of healthcare providers to recognize cardiovascular illness, researchers and experts have devised a variety of clever ways.
 OBJECTIVES: The goal of this research was to propose a robust strategy for cardiac issue prediction utilizing machine learning methods. The healthcare industry generates a massive quantity of data and machine learning has proved effective in making decisions and generating predictions with this data. 
 METHODS: Al has been exhibited to be useful in helping with forecast and decision-production because of the tremendous measure of information made by the medical services a 20 Few explorers have inspected the capability of Al to figure out heart disease. In this article, we suggest a creative strategy. to improve the exactness of cardiovascular sickness forecasts by finding basic highlights utilizing Al systems.
 CONCLUSION: There is a lot of promise and possibility in using machine learning techniques to forecast cardiac disease. By means of examining a range of datasets and applying multiple machine-learning methods. Alongside various element blends and not able arrangement procedures, the expectation model is presented. We accomplish a better exhibition level with a Crossbreed Irregular Woods, with a Direct Model as our coronary illness forecast model.
Tropical Cyclone Detection and Tracking Using YOLOv8 Algorithm Saroja Kumar Rout, Kottu Santosh Kumar, Ruchismita Sahu, Shekharesh Barik, Samarendra Pradhan 2024 1st International Conference on Cognitive Green and Ubiquitous Computing IC Cgu 2024, 2024
