Volume 11 Number 1 June 2024
|
1 |
Navigating Heterogeneity: A Proactive Approach to Handover Management
Abstract - Emerging technologies rely on rapid advancements to boost connectivity. 5G stands out as a crucial enabler for a connected future. However, user mobility poses challenges for seamless handover management. Traditional methods, like structure-based cell measurements, require frequent measurement intervals, impacting performance. The proliferation of ultra-dense small cells (UDSC) alongside macro-cells forms heterogeneous networks (HetNets), leading to increased handovers and radio link failures. Effective mobility management is vital for self-organizing networks to enhance performance. A proposed solution aims to minimize frequent handovers and handover failures. Simulation results comparing 4G and 5G networks demonstrate significant reductions in handover ping-pongs and failures compared to existing literature. |
|
2 |
QoS Aware Optimization of Vertical Handover Decision using Cuckoo Search Algorithm in Heterogeneous Wireless Network
Abstract-The new wireless technology creates a unified worldwide network, offering a variety of services to users worldwide. Its primary aim is to enable mobile users to move effortlessly between different parts of the network. Heterogeneous Wireless Networks (HWNs) enable users with international devices to access the necessary services wherever they are located. Wi-Fi (Wireless Fidelity), Wi-Max (Worldwide Interoperability for Microwave Access), UMTS (Universal Mobile Telecommunication Service), and WLAN (Wireless Local Area Network) are just a few examples of the Radio Access Technologies (RATs) that are included in HWNs. Seamless mobility and the transfer of active calls are facilitated through Vertical Handover (VHO) mechanisms. While ensuring seamless mobility across diverse wireless networks remains a crucial area of research, designing an effective vertical handover decision algorithm is paramount. This technique enables multi-mode terminals to choose the most suitable network among the selections provided. Existing research has introduced a vertical handover decision algorithm for multi-mode terminals employing the Modified Topsis Method, yet it overlooks the aspect of seamless mobility support. Hence the intra-transaction and inter-transaction are hard to focus on achieving the QoS-based decision making. So, in this work, the VHO is considered for the service of the WLAN cellular network. The newly developed Cuckoo search optimization technique (CSO) is employed for the VHO in HWN. Enhancing mobility management entails prioritizing handover (HO) procedures, which are pivotal for addressing mobility-related challenges. A CSO technique has been devised to enhance service quality for end-users, focusing on aspects such as call drop probability, handover delay, jitter, end-to-end delay and throughput. Heterogeneous wireless networks (HWN) are leveraged, taking into account network and achieve superior Quality of Service (QoS). |
|
3 |
Greedy-Based Triclustering of Genetic Algorithm Using Computational Intelligence Techniques (CIT)
Abstract- Triclustering is a popular data mining technique for three dimensional data (3D) analysis. It can reveal hidden and unknown interesting patterns in a 3D data. The main source of 3D data is microarray data which has wide scope in the bioinformatics field especially in drug analysis. Genetic algorithms is the traditional and robust bioinspired optimization technique for data mining tasks. In this work, GA is used to extract tricluster from 3D microarray data efficiently by exploring a large solution space and find highly coherent triclusters. The proposed new strategy for triclustering namely Correlation-Based Triclustering Using Genetic Algorithm (CorrTriGA) that combines the power of GA with a greedy strategy to overcome the shortcomings of poor convergence of GA approaches. A greedy-based technique is specifically incorporated into the GA framework to improve the search process. Fuzzy logic, neural networks, and swarm intelligence are types of computational intelligence (CIDs) that improve tricluster identification process by making the evolutionary process more dynamic. The experimental results carried out on the CDC15 Database to assess the performance of the proposed work. It is shown that CorrTriGA has ability to extract the higher volume coherent tricluster. |
|
4 |
The Diabetes Oracle: Insights from Machine Learning and Predictive Analytics
Abstract- Diabetes is a significant health disorder with potentially severe implications for daily life, increasing the risk of various complications. According to a statistical report from 2045, an alarming estimate of around 135 million individuals may be affected by diabetes, underscoring the magnitude of the issue. Early prediction of diabetes based on symptoms is crucial for effective management. To this end, numerous computational algorithms are employed for early disease diagnosis. The classification of diabetes is done using machine learning techniques such as Naïve Bayes (NB), Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), Ada Boosting (AB), K-Nearest Neighbors (KNN), Gradient Boosting (GB), and Logistic Regression (LR). Through evaluation measures, these methods are systematically compared. Experimental findings reveal that the Gradient Boosting and Ada Boosting methods achieve the highest accuracy of 98%, surpassing other techniques. |
|
5 |
SYSTEMIC UNDERSTANDING OF RISKS AND EXTERNALITIES OF ARTIFICIAL INTELLIGENCE IN AGRICULTURE
Abstract- For over a century, the primary means of boosting agricultural output has been through technological innovation. Novel plant species and synthetic nutrient-management formulas intensification as well as the security of food and nourishment. Rapid plant phenotyping, farmland monitoring, in situ soil composition assessment, disease diagnosis and surveillance, automation and bundling of agrochemical application, weather forecasting, yield prediction, decision support systems (DSS) with real-time agronomic advice, and novel approaches to post-harvest handling and traceability are all potential applications of machine learning (ML) that may be supported and, in some cases, made possible. But agricultural modernity has also brought about ecological deterioration, such as soil erosion and contaminated water and land, which could eventually jeopardise food security. Furthermore, more than 75% of crop genetic variety has been lost as a result of the prioritization of a limited number of plant varieties. |