Downlink Traffic Scheduling in Green Vehicular Roadside Infrastructure

In this paper we consider the problem of scheduling for energy-efficient roadside infrastructure. In certain scenarios, vehicle locations can be predicted with a high degree of accuracy, and this information can be used to reduce downlink infrastructure-to-vehicle energy communication costs. Offline scheduling results are first presented that provide lower bounds on the energy needed to satisfy arriving vehicular communication requirements. We show that the packet-based scheduling case can be formulated as a generalization of the classical single-machine job scheduling problem with a tardiness penalty, which is referred to as $\alpha$–Earliness–Tardiness. A proof is given that shows that even under a simple distance-dependent exponential radio path loss assumption, the problem is NP-complete. The remainder of the paper then focuses on timeslot-based scheduling. We formulate this problem as a Mixed-Integer Linear Program (MILP) that is shown to be solvable in polynomial time using a proposed minimum cost flow graph construction. Three energy-efficient online traffic scheduling algorithms are then introduced for common vehicular scenarios where vehicle position is strongly deterministic. The first, i.e., Greedy Minimum Cost Flow (GMCF), is motivated by our minimum cost flow graph formulation. The other two algorithms have reduced complexity compared with GMCF. The Nearest Fastest Set (NFS) scheduler uses vehicle location and velocity inputs to dynamically schedule communication activity. The Static Scheduler (SS) performs the same task using a simple position-based weighting function. Results from a variety of experiments show that the proposed scheduling algorithms perform well when compared with the energy lower bounds in vehicular situations where path loss has a dominant deterministic component so that energy costs can be estimated. Our results also show that near-optimal results are possible but come with increased computation times compared with our heuristic algorithms.