Revisiting minimum-length scheduling in wireless networks: An algorithmic framework

We consider the problem of constructing the minimum length schedule required to empty a wireless network with queues of given size. In a recent work we have provided new fundamental insights towards its structure and complexity. Motivated by the problem computational complexity, we demonstrate here how a one-size-fits-all optimal algorithm cannot be expected and introduce a framework that decomposes the problem in two core sub-problems: Selecting which subset of wireless links to activate and for how long. This modular approach enables the construction of algorithms that can yield solutions ranging from simple and intuitive to exact optimal. We provide a comprehensive set of design strategies and results to elucidate how different combinations within the framework modules can be used to approach optimality.