Content caching is recognized as a promising solution to release the heavy burden of backhaul links and decrease the content transmission latency in Fog radio access networks (Fog-RANs). However, the content caching design is still a challenging problem with considering the user request patterns, the content delivery strategies, and the limited caching capacity. Recommendation has the capability of reshaping users’ content requests for further prompting caching gain. The joint recommendation, caching, beamforming holds the potential to improve the system performance of Fog-RANs. In this paper, a joint recommendation, caching, and beamforming scheme is proposed for multi-cell multi-antenna recommendation aware Fog-RANs. Aiming at minimizing the content transmission latency, we formulate a joint recommendation, caching, and beamforming optimization problem. The minimization problem is a very challenging two-timescale mixed integer nonlinear programming problem, which is hard to solve in general. By exploring structural properties of the problem, we propose an alternative optimization algorithm with low complexity through decomposing the original problem into three sub-problems. Extensive simulations show that our proposed method can significantly reduce the content transmission delay.

Mixed-Timescale Caching and Beamforming in Content Recommendation Aware Fog-RAN: A Latency Perspective

Caching popular contents at the network edge has been considered as a promising enabler to relieve the pressure on networks due to the fact that a substantial portion of global data traffic is repeatedly requested by many subscribers and thus redundantly generated. Recommendation, on the other hand, has attracted spiraling attention for its capability of reshaping users’ contents demand patterns. In this paper, we examine the practicability of recommendation in boosting the gains of edge caching with uncharted users’ feature information. To this end, we first characterize the average system cost for a generic network model, disclosing its dependence on the recommendation and caching strategies. Then, we formulate the joint caching and recommendation decision oriented cost minimization problem, taking the constraints on each content provider’s cache capacity budget, each individual user’s recommendation size and recommendation quality into account. However, the implicit information regarding users’ preference makes the problem inextricable. To address this issue, a versatile long short term memory (LSTM) network assisted prediction paradigm is proposed to attain the preference schema of users with the assistance of their historical behavior data. Based on that, we rigorously prove the NP-hardness of obtaining the optimal recommendation and caching policies that jointly minimize the system cost. Therewith, an iterative suboptimal algorithm is developed, which has provable polynomial time complexity and convergence guarantee. Extensive simulation results validate the effectiveness of our proposed LSTM enabled feature information prediction approach and the convergence performance of the devised joint decision making methodology. In addition, it is shown that the proposed scheme outperforms numerous benchmarks significantly.

Towards Cost Minimization for Wireless Caching Networks With Recommendation and Uncharted Users’ Feature Information

To maintain reliability of content-oriented wireless caching networks (CWCNs), repair mechanism is of necessity to be considered due to the natural that storage entities are individually unreliable and thus subject to failure on account of hardware error, network congestion or software updating. Meanwhile, recommendation is tunable for edge caching performance improvement. In this paper, we study the revenue maximization problem for CWCNs with both repair and recommendation considerations. The formulated problem is an integer non-convex and non-linear problem, and thus is difficult to be solved. The difficulties are intrinsically derived from the implicit weighted sum costs (WSCs) as regards storage and repair of each content and the coupling among the Boolean variables. For the sake of analytical tractability, a two-step methodology is developed. Specifically, we first explore the optimal storage and repair amount among the content providers to minimize the WSCs in terms of successfully fixing any occurred data corruption for the stored contents. Thereof, an explicit instance is provided to show how the contents can be coded, stored and then repaired in our network given that an error occurs. Based on the obtained storage and repair amount vectors, we solve the resultant joint caching and recommendation decision making problem (DMP). To be more specific, we decouple the DMP into a pair of subproblems, namely a cache placement and a recommendation optimization subproblems. For each subproblem, a globally optimal and a time-efficient suboptimal solutions are developed, respectively. Later, a versatile iterative paradigm is devised to do the decision making jointly. The convergence performance and the complexity analysis of the proposed algorithms are rigorously analyzed. Numerical results confirm the convergence performance of our iterative algorithms and illustrate their revenue improvements compared to various baseline schemes.

Revenue Maximization for Content-Oriented Wireless Caching Networks (CWCNs) With Repair and Recommendation Considerations

The cache-enabled multicast content pushing, which multicasts the content items to multiple users and caches them until requested, is a promising technique to alleviate the heavy network load by enhancing the traffic offloading. This, in turn, has called for the optimization of content pushing strategy while considering both the transmission and caching resources, which jointly result in the complicated coupling among pushing decisions and lead to high computational complexity. Unlike most existing approaches which simplify the pushing problem via bypassing the complicated coupling, in this paper, we propose a multicast content pushing strategy to maximize the offloaded traffic with the cost on content caching based on structured deep learning. Specifically, we design the convolution stage to extract the spatio-temporal correlations of one content item between different pushing decisions, and construct the fully-connected stage to capture the spatial coupling among the decisions of pushing different content items to different user devices. Moreover, to address the absence of the ground truth on multicast content pushing, we relax the transmission constraint to derive a performance upper bound for guiding the training direction. This relaxed problem is solved based on dynamic programming in a bottom-up manner. Compared to the state-of-the-art baselines including both the traditional model-based and the general neural network-based strategies, the proposed pushing strategy achieves significant performance gain in both the random-generated dataset and the real LastFM dataset. In addition, it is also shown that the proposed strategy is robust to the uncertainty of user request information.

Cache-Enabled Multicast Content Pushing With Structured Deep Learning

As user preferences can be influenced by the recommendation system, it has been shown that the joint recommendation and caching policy design can significantly improve the caching networks where caching is at the BSs. However, whether and how the joint recommendation and caching policy design can provide benefits to cache-aided device-to-device (D2D) networks have not been well-understood. This paper thus contributes in this direction by modeling the offloading probability of the cache-aided D2D network and proposing a social-aware joint recommendation caching policy design. Specifically, considering the preferences and social relationship of users as well as the caching and recommendation policies of the network, we formulate an offloading probability optimization problem which is non-convex. Then, an iterative algorithm with monotonicity and convergence property is proposed to solve the problem. By simulations, we show that the proposed joint recommendation and caching policy design can significantly outperform designs that only optimize the caching policy and other reference designs.

Socially-Aware Joint Recommendation and Caching Policy Design in Wireless D2D Networks

Proactive caching is recognized as a promising approach to handle the rapid growth of data traffic, thereby attracting much attention recently. As a key performance metric of caching, the hit ratio is determined by demand probabilities of users for content items and caching decisions. Because the recommendation system is capable of shaping user demands, the joint caching and recommendation holds the potential of improving the hit ratio substantially. In this paper, joint pushing and recommendation (JPR) schemes are presented for multiuser multiple-input single-output (MISO) systems, in which content items are pushed over MISO downlinks with multicast beamforming. Aiming at maximizing the effective throughput, we formulate a multi-stage stochastic programming problem under the constraints of transmit power and quality of experience (QoE). Since the formulated problem is intractable, suboptimal online JPR policies are presented based on the convex–concave procedure and branch-and-bound methods. Simulations show that presented JPR policies are capable of attaining significant effective throughput gains.

Content Pushing Over Multiuser MISO Downlinks With Multicast Beamforming and Recommendation: A Cross-Layer Approach

Proactive caching holds the promise of coping with the explosively increasing traffic demands of next generation mobile networks. However, it may also incur extra pushing and storage costs for base stations (BSs). By taking these costs into account, we are interested in maximizing the average profit for the BS. A joint pushing and caching (JPC) approach is investigated to determine whether and when to push and how long a file should be cached at the user’s buffer. More specifically, we present a dynamic programming based JPC for the BS that knows the distributions of user’s request times for content files. When user’s request time distributions are unknown in priori, we adopt a reinforcement learning algorithm to achieve high bandwidth and storage efficiency.

Bandwidth and Storage Efficient Caching Based on Dynamic Programming and Reinforcement Learning

Joint pushing and caching (JPC) holds the promise of achieving high spectral and energy efficiency in future networks, thereby attracting much attention most recently. However, few attention has been paid to the content consumption process that describes how a content item leaves the receiver buffer after the request time. Such a process bounds the amount of data that have arrived in the buffer, thereby constraining the content transmission. By taking content consumption processes into account, this paper aims to present optimal JPC policies that minimize the transmission energy under receiver buffer size and content consumption constraints. In particular, two common types of content consumption processes, instantaneous reading and sequential reading, along with two content transmission strategies, orthogonal multiple access (OMA) and non-orthogonal multiple access (NOMA), are respectively considered. Energy minimization problems are first formulated as variational problems based on the network calculus approach, and are then equivalently transformed into convex optimization problems in order to reduce the computational complexity. It is shown that JPC policies are capable of significantly reducing the energy consumption when receiver buffer sizes are large. Furthermore, the NOMA based JPC achieves higher energy efficiency than the OMA based JPC by effectively exploiting physical layer multicasting opportunities.

Zhiyuan Lin

Papers

Content Pushing Over Multiuser MISO Downlinks With Multicast Beamforming and Recommendation: A Cross-Layer Approach

Bandwidth and Storage Efficient Caching Based on Dynamic Programming and Reinforcement Learning

Energy-Efficient Pushing With Content Consumption Constraints: A Network Calculus Approach