Data offloading based on Device-to-Device (D2D) can support congestion-prone cellular networks in the face of traffic growth. In this paper, we tackle the design of offloading systems that are aware of the energy limitation of the complementary D2D network. We propose a novel heuristic scheme, named HERRA, using a parametric three-stage method that includes possible variations on the employed strategies in each stage. Performance evaluation, using network simulations in our extended NS-3, shows that HERRA outperforms the related work in the matter of convergence time. As a result of massive speedups, up to six orders of magnitude, HERRA scales very well in denser topologies at the price of having some performance gaps, particularly in terms of packet loss.

In this paper, we propose a novel 3D Underwater Wireless Sensor Network Deployment scheme for solid detection in rivers. Our objective is to minimize the number of deployed underwater sensors within a target field installation while ensuring i) the required Quality of Monitoring (QoM) (i.e., differentiated probabilistic detection) and ii) wireless network connectivity. To generate the best topology, we propose a novel deployment heuristic, named 3D-UWSN-Deploy, based on a subcube tessellation of the monitored field installation and a mixed integer linear program optimization. To gauge the effectiveness of 3D-UWSN-Deploy, we compare it with the most prominent related strategies. Simulation results show that our proposal is scalable and obtains the best performance in terms of cost deployment, quality of monitoring and connectivity.

We reformulate branch-and-bound feature selection employing L/sub /spl infin// or particular L/sub p/ metrics, as mixed-integer linear programming (MILP) problems, affording convenience of widely available MILP solvers. These formulations offer direct influence over individual pairwise interclass margins, which is useful for feature selection in multiclass settings.