Utility Maximization for Wireless Content Caching Networks With Diversified Recommendation

In this work, we study the utility maximization problem for wireless content caching networks with diversified recommendation. To make this happen, we first give the definition of recommendation diversity and investigate its effect on system's utility. Thereafter, we explicitly express the utility function of the system. These analyses enable us to formulate the maximization problem with the cooperation of various practical constraints, such as the recommendation quality and quantity per user, and the cache capacity budget at the base station. We rigorously prove the NP-hardness of the optimization problem and then propose an efficient algorithm to determine the recommendation set per user. Moreover, it is shown that the cache placement subproblem can be optimally solved by the dynamic programming algorithm. As last, we alternatively optimize these two types of variables until they converge. Simulation results show the convergence performance of our devised algorithm and its superiority in terms of utility, diversity, and cache hit rate (CHR) when compared to extensive benchmark schemes.