User Association for Cache Enabled Millimeter Wave Ultra Dense Downlink Networks

User association and coverage analysis are both important for cache enabled dense millimeter-wave (mmWave) downlink communication networks. In this paper, we consider a scenario with two types of access point (AP), i.e., many mmWave base stations (BSs) and a mmWave transmitter following the Poisson point process distribution and uniform distribution on a two-dimensional plane, respectively. Given the most popular caching scheme, we propose an user association scheme adaptive to content placement, and derive its association probability and successful transmission probability (STP). Simulation results verify the analytical results and indicate many useful insights. The association probability grows slowly after the BS density reaches a certain level and it is affected by the transmitting power, especially the cache size of different APs and the backhaul capacity of BSs, transmit power, cache size of different APs and the antenna array size all affect STP, and the antenna array size at the BSs has larger effect. BS power should be carefully adjusted to achieve a tradeoff between BS interference and multiple-BS diversity for good STP performance.