Joint Bandwidth Allocation and Path Selection in WANs with Path Cardinality Constraints

In this paper, we study the joint bandwidth allocation and path selection problem, which is an extension of well-known network utility maximization (NUM) via solving a multi-objective minimization problem under cardinality constraints. Specifically, such formulation captures various types objectives including proportional fairness, average delay, as well load balancing. addition, in order to handle "unsplittable flows", constraints are added, making resulting optimization quite challenging solve due intrinsic non-smoothness nonconvexity. Almost all existing works deal with using relaxation techniques transform it into convex problem. However, provide novel solution framework based on linearized alternating direction method multipliers (LADMM) split original coupling terms several subproblems. We then derive that these subproblems, albeit nonconvex nonsmooth, actually simple easy implement, can be independent interest. Under some mild assumptions, prove any limiting point generated sequence proposed algorithm stationary point. Numerical simulations performed demonstrate advantages our compared base-lines.