Volume: 19 Issue-01 (January-June) 2024

---

OPTIMIZING DATA PLACEMENT AND CLASSIFICATION: A HYBRID APPROACH WITH GENETIC ALGORITHMS AND PARTICLE SWARM OPTIMIZATION

ABSTRACT

This paper proposes an Effective Data Emplacement and Classification (EDEC) approach for large-scale streaming data applications in heterogeneous cloud environments. It introduces a clustering model utilizing map and reduce functions to efficiently provision and classify data across resources. The approach leverages distributed computing to handle large volumes of unstructured data, offering a cost-effective solution. To address load imbalance, the model incorporates heterogeneity-aware scheduling mechanisms. The methodology involves MapReduce cluster constraints, dynamic imbalance data processing, task clustering, and adaptive task tuning. Genetic algorithms and particle swarm optimization aid in task scheduling and regrouping for improved performance. Additionally, bipartite graph modeling facilitates efficient resource allocation. The proposed approach is evaluated using Hadoop and demonstrates enhanced scalability and efficiency in processing large datasets.

FULL TEXT

PDF