Optimasi Cloud Task Scheduling Menggunakan Bat Algorithm Dengan Opposition-Based Learning (BA-OBL) Pada Cloud Environment

Rahmat, Abdul Zaki Syahrul (2025) Optimasi Cloud Task Scheduling Menggunakan Bat Algorithm Dengan Opposition-Based Learning (BA-OBL) Pada Cloud Environment. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027211020-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Salah satu tantangan utama dalam komputasi awan adalah task scheduling, yang bertujuan untuk mengoptimalkan alokasi tugas pada sumber daya yang tersedia. Penelitian ini mengusulkan pengembangan dan integrasi Bat Algorithm (BA) dengan pendekatan OppositionBased Learning (OBL) untuk meningkatkan efisiensi dan efektivitas dalam task scheduling. BA, yang terinspirasi oleh proses ekolokasi kelelawar, menawarkan metode inovatif dalam pencarian solusi, sementara OBL mempercepat eksplorasi ruang pencarian dengan mempertimbangkan solusi kebalikan.
Untuk memvalidasi performa algoritma yang diusulkan, penelitian ini dilakukan dalam dua lingkungan berbeda, yaitu lingkungan simulasi menggunakan framework CloudSim dengan dataset Simple Random, Stratified Random, dan The San Diego Supercomputer Center (SDSC), serta implementasi pada lingkungan nyata berskala kecil menggunakan dataset Real Environment (RE) yang dibangun menggunakan kontainer Docker. Hasil pengujian menunjukkan bahwa algoritma BA-OBL secara konsisten mengungguli algoritma pembanding yaitu Genetic Algorithm (GA), Particle Swarm Optimization (PSO), dan BA tanpa OBL di kedua lingkungan pengujian.
Pada lingkungan simulasi, BA-OBL berhasil menurunkan makespan hingga 38,47%, serta meningkatkan throughput sebesar 63,81% dan resource utilization hingga 82,79% dibandingkan GA dan PSO. Pada implementasi nyata, BA-OBL juga menunjukkan keunggulan dengan penurunan makespan sebesar 14,85% dan imbalance degree hingga 75,25%. Meskipun demikian, tercatat sedikit peningkatan pada average execution time. Hasil ini membuktikan bahwa integrasi OBL secara signifikan meningkatkan performa BA, menjadikannya solusi yang sangat efektif dan aplikatif untuk optimasi task scheduling serta berkontribusi pada pengembangan algoritma metaheuristik yang lebih optimal untuk meningkatkan pemanfaatan sumber daya secara nyata dalam lingkungan komputasi awan.
==================================================================================================================================
One of the main challenges in cloud computing is task scheduling, which aims to optimize the allocation of tasks to available resources. This research proposes the development and integration of the Bat Algorithm (BA) with an Opposition-Based Learning (OBL) approach to enhance efficiency and effectiveness in task scheduling. BA, inspired by the echolocation process of bats, offers an innovative method for solution searching, while OBL accelerates the exploration of the search space by considering opposite solutions.
To validate the performance of the proposed algorithm, this research was conducted in two different environments: a simulation environment using the CloudSim framework with Simple Random, Stratified Random, and The San Diego Supercomputer Center (SDSC) datasets, and a small-scale real-world implementation using a Real Environment (RE) dataset built using Docker containers. The test results show that the BA-OBL algorithm consistently outperforms the comparison algorithms, namely Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and standard BA, in both testing environments.
In the simulation environment, BA-OBL successfully reduced the makespan by up to 38.47%, while increasing throughput by 63.81% and resource utilization by up to 82.79% compared to GA and PSO. In the real-world implementation, BA-OBL also demonstrated superiority with a 14.85% reduction in makespan and up to a 75.25% reduction in imbalance degree. However, a slight increase in the average execution time was noted. These results prove that the integration of OBL significantly enhances the performance of BA, making it a highly effective and applicable solution for task scheduling optimization. This work contributes to the development of more optimal metaheuristic algorithms for improving resource utilization in real-world cloud computing environments.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Komputasi Awan, Cloud Computing, CloudSim, Docker, Task Scheduling, Bat Algoritm, Opposition-Based Learning.
Subjects:	Q Science > QA Mathematics > QA76.585 Cloud computing. Mobile computing.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Abdul Zaki Syahrul Rahmat
Date Deposited:	30 Jul 2025 08:51
Last Modified:	30 Jul 2025 08:51
URI:	http://repository.its.ac.id/id/eprint/123257

Actions (login required)

View Item