Data-Driven Optimized Slice Activation in Multi-Tenant 5G Networks

Network slicing to support multi-tenancy plays a key role in improving the performance of 5G networks. In this paper, we propose a reservation-based framework to activate a subset of network slices with the goal of maximizing the expected utilities of tenants and minimizing the expected outage of downlinks. Based on the statistics from users and channels and also considering the expected utility from serving users of a slice and the activation cost, we formulate a sparse combinatorial optimization problem to activate network slices to provide services to users. The formulated optimization is mixed-binary, non-convex and difficult to solve. We propose a Frank-Wolfe algorithm to iteratively solve a sequence of relaxed convex approximations of the original problem. Simulation results demonstrate the performance of our approach relative to optimal solutions.