Cooperative Cache-Aware Recommendation System for Multiple Internet Content Providers

Joint optimization of caching and recommendation is expected to achieve a beneficial tradeoff between caching efficiency and users' quality of experience (QoE). To this end, we study a cooperative cache-aware recommendation system (CCARS) with multiple Internet content providers (ICPs) considering content sharing among them. We first analytically derive the cache hit ratio by modeling users' preference distribution and the impact of recommendation. We then formulate a problem of maximizing the cache hit ratio by optimizing the caching placement and the recommendation policy. Finally, a heuristic algorithm with low computational complexity is proposed to solve the formulated problem. The effectiveness of our algorithm is validated and the advantage with respect to the independent cache-aware recommendation system (ICARS) for each ICP is demonstrated. Numerical results show that the cache hit ratio benefits from the increased cache size at base station (BS) and the augmented content sharing among ICPs.