Parallel genetic algorithm based optimal fusion in sensor networks
Nithya Gnanapandithan, Balasubramaniam Natarajan · 2006
In this paper, we propose a Parallel Genetic Algo- rithm (PGA) for optimizing the performance of a parallel decen- tralized sensor network. The metric used for the optimization is the probability of global detection error. The Parallel Genetic Algorithm simultaneously optimizes both the fusion rule and the local decision rules. We show that our approach provides results comparable to those obtained by using a GA and gradient- based algorithm from previous work by Aldosari and Moura, with reduced complexity. We consider both the cases of identical (homogeneous) and non-identical (heterogeneous) sensors and demonstrate that our algorithm converges to the same optimal solution in both cases. We also discuss the effect of the quality of the initial solution on the convergence of the PGA.