Paper Abstract and Keywords |
Presentation |
2024-03-01 11:35
Application of a Deep Reinforcement Learning Algorithm to Virtual Machine Migration Control in Multi-Stage Information Processing Systems Yuki Kojitani (Okayama Univ.), Kazutoshi Nakane (Nagoya Univ.), Yuya Tarutani (Okayama Univ.), Celimuge Wu (UEC), Yusheng Ji (NII), Tokumi Yokohira (Okayama Univ.), Tutomu Murase (Nagoya Univ.), Yukinobu Fukushima (Okayama Univ.) IN2023-87 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
This paper tackles a virtual machine (VM) migration control problem to maximize the progress (accuracy) of information processing tasks in multi-stage information processing systems. The conventional methods for this problem (e.g., VM sweeping method and load balancing method) are effective only for specific situations, such as when the system load is high. In this paper, in order to achieve high accuracy in various situations, we propose a VM migration method using a deep reinforcement learning algorithm. It is difficult to directly apply a deep reinforcement learning algorithm to the VM migration control problem because the size of the action space of an agent dynamically changes according to the number of VMs in the system while the size of the action space is fixed in deep reinforcement learning algorithms. Therefore, the proposed method divides the VM migration control problem into two problems: the problem of determining only the VM distribution and the problem of determining the locations of all the VMs so that it follows the determined VM distribution. The former problem is solved by a deep reinforcement learning algorithm, and the latter problem is solved by a heuristic method. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Multi-stage information processing system / VM migration control / Deep reinforcement learning / DDPG (Deep Deterministic Policy Gradient) / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 123, no. 398, IN2023-87, pp. 130-135, Feb. 2024. |
Paper # |
IN2023-87 |
Date of Issue |
2024-02-22 (IN) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
IN2023-87 |
|