Hide/Show Apps

Bio-inspired bandwidth packing

KORUK, Talha
Onur, Ertan
Sharing channel capacity among multi-rate sensors or synonymously packing the bandwidth while satisfying quality of service requirements stays as an important challenge. We present bio-inspired solutions to this problem by reducing it to the NP-hard multiple-choice knapsack problem. We employ various bio-inspired population-based meta-heuristics to allocate capacity to the requesting nodes in a sensor network. In this work-in-progress paper, we present the controlled lab experiments for determining the capacity of a wireless channel and then discuss the feasibility of meta-heuristic solutions. The runtime and closeness to the optimal solutions results are presented and discussed. Artificial bee colony optimisation provides the fastest solution although the convergence rate per generation is slower.