作者: Jeffrey Wildman , Yusuf Osmanlioglu , Steven Weber , Ali Shokoufandeh
关键词: Context (language use) 、 Linearization 、 Mathematical optimization 、 Tree (graph theory) 、 Computer science 、 Quadratic equation 、 Linear approximation 、 Cellular network 、 Quadratic programming
摘要: We study downlink delay minimization within the context of cellular user association policies that map mobile users to base stations. note minimum problem fits a broader class network utility maximization and can be posed as non-convex quadratic program. This non-convexity motivates split objective function captures original problem's inherent tradeoff: with station provides highest signal-to-interference-plus-noise ratio (SINR) vs. is least congested. find split-term formulation amenable linearization by embedding stations in hierarchically well-separated tree (HST), which offers linear approximation constant distortion. provide numerical comparison several formulations appropriate optimization parameter selection, reformulation produces sum delays are close maximization. also comment on more difficult when idle (those without associated users) deactivated.