Multi-Armed Bandit Learning for Cache Content Placement in Vehicular Social Networks

Bitaghsir, Saeid Akhavan; Dadlani, Aresh; Borhani, Mohammad; Khonsari, Ahmad

doi:10.1109/lcomm.2019.2941482

Cited by 15 publications

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Okada et al [ 37 ] calculate the communication volume between users and caching nodes from the access probability and the number of communication hops, and select the caching location based on the communication volume. On the Internet of Vehicles (IoV), Bitaghsir et al [ 38 ] proposed a caching placement algorithm based on Multi-Armed Bandit Learning, which selects the content to be cached in the roadside unit (RSU) according to the content popularity, and then uses the user social characteristics to select the optimal caching path. This algorithm effectively reduces the load of each caching node.…”

Section: Placement Optimizationmentioning

confidence: 99%

A Comprehensive Review on Edge Caching from the Perspective of Total Process: Placement, Policy and Delivery

Fan

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

With the explosive growth of smart devices and mobile applications, mobile core networks face the challenge of exponential growth in traffic and computing demand. Edge caching is one of the most promising solutions to the problem. The main purpose of edge caching is to place popular content that users need at the edge of the network, borrow free space to reduce user waiting time, and lighten the network load by reducing the amount of duplicate data. Due to the promising advantages of edge caching, there have been many efforts motivated by this topic. In this paper, we have done an extensive survey on the existing work from our own perspectives. Distinguished from the existing review articles, our work not only investigates the latest articles in this area, but more importantly, covers all the researches of the total process of edge caching from caching placement optimization, policy design, to the content delivery process. In particular, we discuss the benefits of caching placement optimization from the perspective of different stakeholders, detail the delivery process, and conduct in-depth discussions from the five phases, i.e., requested content analysis, user model analysis, content retrieval, delivery, and update. Finally, we put forward several challenges and potential future directions, and hope to bring some ideas for the follow-up researches in this area.

show abstract

Section: Placement Optimizationmentioning

confidence: 99%

A Comprehensive Review on Edge Caching from the Perspective of Total Process: Placement, Policy and Delivery

Fan

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…The influence of social relations on task unloading in a vehicle cloud network was studied to improve the link stability, thus addressing the vehicular QoS demand. Bitaghsir et al [21] proposed a resource allocation algorithm based on the robber 2 study that considered the impact of centrality as a social attribute on content delivery between vehicles to maximize the probability of requesting vehicles to download cache data and improve the efficiency of content delivery successfully.…”

Section: Introductionmentioning

confidence: 99%

Resource Allocation Strategy Based On Tripartite Graph in Vehicular Social Networks

Zhang

Zhou

2023

IEEE Trans. Netw. Sci. Eng.

View full text Add to dashboard Cite

Owing to the limited spectrum of vehicular social networks and the fragmented distribution of caching resources, problems such as low vehicle data transmission rate, poor service quality, and low service content delivery efficiency of streaming media are encountered. To solve these problems, we propose a resource allocation strategy based on tripartite graph in vehicular social networks (RATG). This strategy establishes a mobile vehicular social networks model based on vehicle mobility and social similarity. To maximize the content delivery efficiency and transmission rate, one-to-one stable matching based on mobile social connections and one-to-one stable matching based on channel state are established. The proposed strategy incorporates a heuristic optimization algorithm based on the maximum spanning tree, such that the matching results of the two tripartite graphs can reach the approximate global optimal simultaneously and further improve the service quality of vehicles. The simulation results show that the content delivery efficiency of the RATG strategy is 6.05%, 5.00%, and 10.64% higher than those of the randomized resources allocation strategy, local search strategy, and local ratio strategy, respectively, and the average transmission rate is at least 2.7% higher for the RATG strategy, relative to those of the other strategies.

show abstract

“…These studies seldom consider the uncertainty of users' behaviours, so this paper introduces an online learning method called multi-armed bandits (MAB) to solve the problem. MAB has shown effectiveness and merit in air conditioning demand aggregation [16] and many other sequential decisionmaking problems containing uncertain/unknown behavioural factors [17][18][19][20][21][22][23][24][25][26][27]. In reference [28], an adversarial MAB framework is applied to learn the signal response of thermal control loads for demand response in real-time.…”

Section: Introductionmentioning

confidence: 99%

A dynamic distributed energy storage control strategy for providing primary frequency regulation using multi‐armed bandits method

Sun

Zhao

Zhang

et al. 2021

IET Generation Trans & Dist

View full text Add to dashboard Cite

Maintaining frequency stability is a crucial but challenging task for the stable operation of a power system. The distributed energy storage (DES) can charge or discharge for both upward and downward frequency regulation, exploring and effectively using the regulation capabilities will provide a strong backup for frequency regulation. Here, a dynamic DES control strategy for providing primary frequency regulation is proposed. The different behaviours of storage owners are considered when they respond to the regulation requests from the aggregator. This kind of uncertainty would lead to a mismatch between the final aggregation result and the expected target. Hence, the multi-armed bandit approach is applied to learn users' response behaviour and select the optimal set of users to mitigate the mismatch. Case studies on the IEEE RTS 24-bus system demonstrate that the proposed method's mismatch between the actual aggregation result and the regulation target is less than half as much as the conventional method, and it can restore the frequency 20 events earlier than the traditional method.

show abstract

Multi-Armed Bandit Learning for Cache Content Placement in Vehicular Social Networks

Cited by 15 publications

References 13 publications

A Comprehensive Review on Edge Caching from the Perspective of Total Process: Placement, Policy and Delivery

A Comprehensive Review on Edge Caching from the Perspective of Total Process: Placement, Policy and Delivery

Resource Allocation Strategy Based On Tripartite Graph in Vehicular Social Networks

A dynamic distributed energy storage control strategy for providing primary frequency regulation using multi‐armed bandits method

Contact Info

Product

Resources

About