Energy-efficient multi-connectivity enabled user association and downlink power allocation in mmWave networks

Multi-Connectivity enabled User Association (MCUA) is a promising technology to overcome the severe path loss and the high sensitiveness to obstacles of Millimeter Wave (mmWave) signals in ultra-dense mmWave networks. This paper aims to jointly optimize the MCUAs and the downlink Power Allocations (PAs) to maximize the overall Energy Efficiency (EE) of all mmWave Base Stations (mBSs) while guaranteeing the minimum downlink rate. This is a non-convex Mixed-Integer Nonlinear Programming (MINLP) with the form of summation of fractions. To solve it, we first transform the problem into a min-max optimization using the equal-weighted Tchebycheff method to avoid handling the summation of fractions. Next, we relax all binary variables into continuous ones. This relaxed problem is then solved jointly by the Dinkelbach-type method and successive convex approximation so that the upper-bound is obtained. The final mixed-integer solution is obtained by the branch-and-bound method. Numerical results demonstrate the effectiveness of the proposed algorithms.