Cooperative Local Caching under Heterogeneous File Preferences
Local caching is an effective scheme for leveraging the memory of the mobile terminal (MT) and short range communications to save the bandwidth usage and reduce the download delay in the cellular communication system. Specifically, the MTs first cache in their local memories in off-peak hours and then exchange the requested files with each other in the vicinity during peak hours. However, prior works largely overlook MTs' heterogeneity in file preferences and their selfish behaviours. In this paper, we practically categorize the MTs into different interest groups according to the MTs' preferences. Each group of MTs aims to increase the probability of successful file discovery from the neighbouring MTs (from the same or different groups). Hence, we define the groups' utilities as the probability of successfully discovering the file in the neighbouring MTs, which should be maximized by deciding the caching strategies of different groups. By modelling MTs' mobilities as homogeneous Poisson point processes (HPPPs), we analytically characterize MTs' utilities in closed-form. We first consider the fully cooperative case where a centralizer helps all groups to make caching decisions. We formulate the problem as a weighted-sum utility maximization problem, through which the maximum utility trade-offs of different groups are characterized. Next, we study two benchmark cases under selfish caching, namely, partial and no cooperation, with and without inter-group file sharing, respectively. The optimal caching distributions for these two cases are derived. Finally, numerical examples are presented to compare the utilities under different cases and show the effectiveness of the fully cooperative local caching compared to the two benchmark cases.