Energy-Efficient Scheduling in Green Vehicular Infrastructure With Multiple Roadside Units

In this paper, the authors propose low-complexity algorithms for downlink traffic scheduling in green vehicular roadside infrastructure. In multiple roadside unit (RSU) deployments, the energy provisioning of the RSUs may differ, and it is therefore desirable to balance RSU usage from a normalized min-max energy viewpoint. This paper considers both splittable RSU assignment (SRA) and unsplittable RSU asssignment (URA) scheduling. An offline integer linear programming bound is first derived for normalized min-max RSU energy usage. The authors then show that in the SRA case, there is a polynomial complexity 2-approximation bound for the normalized min-max energy schedule. This paper then proposes several online scheduling algorithms. The first is a greedy online algorithm that makes simple RSU selections, followed by minimum-energy time slot assignments. A normalized min-max algorithm is then proposed [2-approximation online algorithm (TOAA)], which is an online version of the 2-approximation bound. Two algorithms are then introduced based on a potential function scheduling approach. The 1-objective algorithm uses an objective based on normalized min-max energy, and the authors show that it has an upper bounded worst-case competitive ratio performance. The 2-objective algorithm uses the same approach but incorporates a total-energy secondary objective as well. Results from a variety of experiments show that the proposed scheduling algorithms perform well. In particular, they find that in the SRA case, the TOAA algorithm performs very close to the lower bound but at the expense of having to reassign time slots whenever a new vehicle arrives. In the URA case, the low-complexity 1-objective algorithm performs better than the others over a wide range of traffic conditions.


  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01567937
  • Record Type: Publication
  • Files: TRIS
  • Created Date: May 19 2015 10:43AM