Cloud Provisioning Menggunakan Genetic Algorithm Dan Artificial Neural Network

Mannuel, Bryan Yehuda (2023) Cloud Provisioning Menggunakan Genetic Algorithm Dan Artificial Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05311940000021-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2025. Download (6MB) | Request a copy

Abstract

Di era modern dimana penggunaan teknologi semakin pesat dan meningkat secara cepat, penggunaan Cloud Computing semakin banyak diminati. Oleh karena itu, diperlukan sebuah sistem manajemen sumber daya yang baik bagi sebuah Cloud Service Provider agar sistem Cloud Computing mereka dapat memanfaatkan kemampuan virtualisasi sumber daya secara maksimal dan meningkatkan tingkat penggunaan sumber daya Cloud. Namun, tantangan terbesar dalam membangun sebuah sistem manajemen sumber daya dalam Cloud Computing adalah mencari algoritma yang bisa memaksimalkan penggunaan sumber daya Cloud.
Untuk bisa mengatasi tantangan tersebut maka diadakanlah penelitian menggunakan algoritma Genetic Algorithm yang terinspirasi dari proses seleksi natural dan implementasi Artificial Neural Network yang didasarkan pada jaringan saraf biologis yang membentuk otak untuk membangun sebuah sistem penjadwalan tugas dan alokasi mesin virtual (VM) untuk memaksimalkan penggunaan sumber daya Cloud.

Penelitian dilakukan dalam dua skenario, dimana untuk skenario pertama akan digunakan Genetic Algorithm saja dan untuk skenario kedua akan digunakan Genetic Algorithm bersamaan dengan Artificial Neural Network. Pada kedua skenario ini, akan dihasilkan peningkatan efisiensi penjadwalan tugas pada tiap iterasi penggunaan Genetic Algorithm ataupun tiap iterasi pelatihan Artificial Neural Network. Iterasi ini akan dilakukan secara terus menerus hingga tidak didapati peningkatan efisiensi yang signifikan atau kondisi terminasi telah dipenuhi. Kedua skenario tersebut kemudian akan dilakukan perbandingan untuk mencari tahu mana sistem Cloud Provisioning yang lebih efisien berdasarkan berbagai macam parameter penilaian.

Hasil penelitian memberikan kesimpulan bahwa implementasi Genetic Algorithm saja bisa menghasilkan tingkat Resource Utilization sekitar 48% hingga 60% dan implementasi Genetic Algorithm bersamaan dengan Artificial Neural Network sekitar 38% hingga 59%. Implementasi Genetic Algorithm saja berguna untuk menghemat penggunaan energi, memaksimalkan Resource Utilization, dan mengurangi Execution Time. Sedangkan implementasi Genetic Algorithm bersamaan dengan Artificial Neural Network bisa digunakan saat memiliki banyak Task dengan ketidakseimbangan data yang besar dan ingin cepat melakukan Cloud Provisioning.
===================================================================================================================================
In this modern era where the use of technology is growing rapidly and increasing rapidly, the use of Cloud Computing is increasingly in demand. Therefore, a good resource management system is needed for a Cloud Service Provider so that their Cloud Computing system can take full advantage of resource virtualization capabilities and increase the level of resource utilization of Cloud resources. However, the biggest challenge in building a resource management system in Cloud Computing is to find an algorithm that can maximize the use of Cloud resources.

To be able to overcome these challenges, this research was conducted using a Genetic Algorithm inspired by the natural selection process and the implementation of an Artificial Neural Network which is based on a biological neural network that forms the brain to build a task scheduling and virtual machine (VM) allocation system to maximize the use of cloud resources.

The research was conducted using two different scenarios, the first of which just used the Genetic Algorithm and the second of which combined it with an Artificial Neural Network. With each iteration of the Genetic Algorithm or Artificial Neural Network training in these two scenarios, task scheduling effectiveness will rise. Until no appreciable improvement in efficiency is identified or the termination conditions have been satisfied, this iteration will continue. After that, the two scenarios will be contrasted to determine whether Cloud Provisioning solution is more effective based on several evaluation parameter.

The results of the study conclude that the implementation of a Genetic Algorithm alone can produce a Resource Utilization level of around 48% to 60% and the implementation of a Genetic Algorithm together with an Artificial Neural Network is around 38% to 59%. Genetic Algorithm implementation alone is useful for saving energy use, maximizing Resource Utilization, and reducing Execution Time. Meanwhile, the implementation of a Genetic Algorithm together with an Artificial Neural Network can be used when you have many Tasks with large data imbalances and want to quickly do Cloud Provisioning.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Artificial Neural Network, Cloud Computing, Genetic Algorithm, Mesin Virtual, Penjadwalan Tugas, Artificial Neural Network, Cloud Computing, Genetic Algorithm, Task Scheduling, Virtual Machine
Subjects:	Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Bryan Yehuda Mannuel
Date Deposited:	02 Feb 2023 06:42
Last Modified:	02 Feb 2023 06:42
URI:	http://repository.its.ac.id/id/eprint/96013

Actions (login required)

View Item