Joint User-Association and Resource-Allocation in Virtualized Wireless Networks

Abstract

In this paper, we consider a down-link transmission of multicell virtualized wireless networks (VWNs) where users of different service providers (slices) within a specific region are served by a set of base stations (BSs) through orthogonal frequency division multiple access (OFDMA). In particular, we develop a joint BS assignment, sub-carrier and power allocation algorithm to maximize the network throughput, while satisfying the minimum required rate of each slice. Under the assumption that each user at each transmission instance can connect to no more than one BS, we introduce the user-association factor (UAF) to represent the joint sub-carrier and BS assignment as the optimization variable vector in the mathematical problem formulation. Sub-carrier reuse is allowed in different cells, but not within one cell. As the proposed optimization problem is inherently non-convex and NP-hard, by applying the successive convex approximation (SCA) and complementary geometric programming (CGP), we develop an efficient two-step iterative approach with low computational complexity to solve the proposed problem. For a given power-allocation, Step 1 derives the optimum userassociation and subsequently, for an obtained user-association,Step 2 find the optimum power-allocation. Simulation results demonstrate that the proposed iterative algorithm outperforms the traditional approach in which each user is assigned to the BS with the largest average value of signal strength, and then, joint sub-carrier and power allocation is obtained for the assigned users of each cell. Especially, for the cell-edge users, simulation results reveal a coverage improvement up to 57% and 71% for uniform and non-uniform users distribution, respectively leading to more reliable transmission and higher spectrum efficiency for VWN.Index Terms-Complementary geometric programming, successive convex approximation, joint user association and resource allocation, virtualized wireless networks.

Dynamic Resource Allocation for Virtualized Wireless Networks in Massive-MIMO-Aided and Fronthaul-Limited C-RAN

“…To solve them efficiently, by applying CGP [12] along with various transformations and convexification approaches, the sequence of lower bound geometric programming based approximation is derived. We refer interested reader about CGP to Appendix A in [13]. The sub-algorithms are described in details in the following subsections.…”

“…The problem for each step is still non-convex and NP-hard. We apply the complementary geometric programming (CGP) and successive approximation approach (SCA) via different relaxation and transformation techniques, the sub-problems in each step are transformed into the geometric programming (GP) counterparts [12], [13], which can be solved efficiently by optimization packages like CVX [14].…”

Joint uplink and downlink delay‐aware resource allocation in C‐RAN

“…We assume that each user can be connected to one RRH at a time slot; hence, C1 ensures that when user k is assigned to RRH m with subcarrier n f , that user would not be assigned by other RRHs m′ with any subcarriers.…”

“…To propose an efficient algorithm, we apply successive convex approximation (SCA) and diverse techniques of relaxation on convexification such as arithmetic‐geometric mean approximation (AGMA) . Following a two‐step iterative algorithm for UAFs assignment and power allocation, we reshape the problem of each step into geometric programming (GP) . Therefore, our problem can be solved efficiently by software optimization package, same as CVX. By simulation results, we demonstrate how optimizing joint uplink and downlink transmission parameters can improve the performance of the network in contrast to the traditional approach where the one‐way delay was considered and disjointly allocated resources for uplink and downlink session.…”

Power-Efficient Resource Allocation in NOMA Virtualized Wireless Networks

“…The objective is to minimize the total transmit power in a VWN, while maintaining the minimum required capacity for each slice. Since the original problem is non-convex and computationally intractable, we use the approach of complementary Geometric programming (CGPA) and arithmetic-geometric mean inequality (AGMA) to convert it into an efficient algorithm [17], [18]. The simulation results demonstrate that NOMA is more power-efficient than OFDMA in various scenarios.…”