A smart cache content update policy based on deep reinforcement learning

This paper proposes a DRL-based cache content update policy in the cache-enabled network to improve the cache hit ratio and reduce the average latency. In contrast to the existing policies, a more practical cache scenario is considered in this work, in which the content requests vary by both time an...

Full description

Bibliographic Details
Main Authors: Li, Lincan, Kwong, Chiew Foong, Liu, Qianyu, Wang, Jing
Format: Article
Language:English
Published: Hindawi Limited 2020
Subjects:
Online Access:https://eprints.nottingham.ac.uk/64083/
_version_ 1848800087985094656
author Li, Lincan
Kwong, Chiew Foong
Liu, Qianyu
Wang, Jing
author_facet Li, Lincan
Kwong, Chiew Foong
Liu, Qianyu
Wang, Jing
author_sort Li, Lincan
building Nottingham Research Data Repository
collection Online Access
description This paper proposes a DRL-based cache content update policy in the cache-enabled network to improve the cache hit ratio and reduce the average latency. In contrast to the existing policies, a more practical cache scenario is considered in this work, in which the content requests vary by both time and location. Considering the constraint of the limited cache capacity, the dynamic content update problem is modeled as a Markov decision process (MDP). Besides that, the deep Q-learning network (DQN) algorithm is utilised to solve the MDP problem. Specifically, the neural network is optimised to approximate the Q value where the training data are chosen from the experience replay memory. The DQN agent derives the optimal policy for the cache decision. Compared with the existing policies, the simulation results show that our proposed policy is 56%-64% improved in terms of the cache hit ratio and 56%-59% decreased in terms of the average latency.
first_indexed 2025-11-14T20:46:00Z
format Article
id nottingham-64083
institution University of Nottingham Malaysia Campus
institution_category Local University
language English
last_indexed 2025-11-14T20:46:00Z
publishDate 2020
publisher Hindawi Limited
recordtype eprints
repository_type Digital Repository
spelling nottingham-640832020-12-18T08:45:11Z https://eprints.nottingham.ac.uk/64083/ A smart cache content update policy based on deep reinforcement learning Li, Lincan Kwong, Chiew Foong Liu, Qianyu Wang, Jing This paper proposes a DRL-based cache content update policy in the cache-enabled network to improve the cache hit ratio and reduce the average latency. In contrast to the existing policies, a more practical cache scenario is considered in this work, in which the content requests vary by both time and location. Considering the constraint of the limited cache capacity, the dynamic content update problem is modeled as a Markov decision process (MDP). Besides that, the deep Q-learning network (DQN) algorithm is utilised to solve the MDP problem. Specifically, the neural network is optimised to approximate the Q value where the training data are chosen from the experience replay memory. The DQN agent derives the optimal policy for the cache decision. Compared with the existing policies, the simulation results show that our proposed policy is 56%-64% improved in terms of the cache hit ratio and 56%-59% decreased in terms of the average latency. Hindawi Limited 2020-11-09 Article PeerReviewed application/pdf en cc_by https://eprints.nottingham.ac.uk/64083/1/A%20Smart%20Cache%20Content%20Update%20Policy%20Based%20on%20Deep%20Reinforcement%20Learning.pdf Li, Lincan, Kwong, Chiew Foong, Liu, Qianyu and Wang, Jing (2020) A smart cache content update policy based on deep reinforcement learning. Wireless Communications and Mobile Computing, 2020 . pp. 1-11. ISSN 1530-8669 cache; deep reinforcement learning (DRL); deep Q-learning network (DQN); content update; cache hit ratio; average latency; spatiotemporally varying http://dx.doi.org/10.1155/2020/8836592 doi:10.1155/2020/8836592 doi:10.1155/2020/8836592
spellingShingle cache; deep reinforcement learning (DRL); deep Q-learning network (DQN); content update; cache hit ratio; average latency; spatiotemporally varying
Li, Lincan
Kwong, Chiew Foong
Liu, Qianyu
Wang, Jing
A smart cache content update policy based on deep reinforcement learning
title A smart cache content update policy based on deep reinforcement learning
title_full A smart cache content update policy based on deep reinforcement learning
title_fullStr A smart cache content update policy based on deep reinforcement learning
title_full_unstemmed A smart cache content update policy based on deep reinforcement learning
title_short A smart cache content update policy based on deep reinforcement learning
title_sort smart cache content update policy based on deep reinforcement learning
topic cache; deep reinforcement learning (DRL); deep Q-learning network (DQN); content update; cache hit ratio; average latency; spatiotemporally varying
url https://eprints.nottingham.ac.uk/64083/
https://eprints.nottingham.ac.uk/64083/
https://eprints.nottingham.ac.uk/64083/