Caching Policy Optimization for D2D Communications by Learning User Preference

TL;DR

This paper proposes a caching policy for D2D communications that learns individual user preferences using probabilistic models, significantly improving offloading performance over traditional popularity-based methods.

Contribution

It introduces a user preference-based caching optimization framework and employs pLSA and EM algorithms for preference prediction, enhancing D2D caching efficiency.

Findings

User preferences can be learned quickly.

Preference-based caching outperforms popularity-based caching.

Significant offloading gain improvement.

Abstract

Cache-enabled device-to-device (D2D) communications can boost network throughput. By pre-downloading contents to local caches of users, the content requested by a user can be transmitted via D2D links by other users in proximity. Prior works optimize the caching policy at users with the knowledge of content popularity, defined as the probability distribution of request for every file in a library from by all users. However, content popularity can not reflect the interest of each individual user and thus popularity-based caching policy may not fully capture the performance gain introduced by caching. In this paper, we optimize caching policy for cache-enabled D2D by learning user preference, defined as the conditional probability distribution of a user's request for a file given that the user sends a request. We first formulate an optimization problem with given user preference to maximize the offloading probability, which is proved as NP-hard, and then provide a greedy algorithm to find the solution. In order to predict the preference of each individual user, we model the user request behavior by probabilistic latent semantic analysis (pLSA), and then apply expectation maximization (EM) algorithm to estimate the model parameters. Simulation results show that the user preference can be learnt quickly. Compared to the popularity-based caching policy, the offloading gain achieved by the proposed policy can be remarkably improved even with predicted user preference.