International Journal of Swarm Intelligence and Evolutionary Computation

Abstract

Baldwinian Learning in Quantum Evolutionary Algorithms for Solving the Fine-Grained Localization Problem in Wireless Sensor Networks

Mahdi Aziz and Mohammad Meybodi

A Local Search (LS) procedure is a search facilitator, giving memetic algorithms a hand to enhance their exploitation ability resulting in converging to higher quality solutions. In this paper, using the LS procedure in the form of Baldwinian Learning (BL) a Memetic Quantum Evolutionary Algorithm (QEA) is proposed for tackling the fine grained localization problem in Wireless sensor networks (WSNs). Since the QEA can be used only for binarydomain problems like the knapsack problem, we utilize the binary-to-real mapping procedure to make it suitable for solving the localization problem in WSNs. To provide good initial positions of sensor nodes, the algorithm employs a Multi-Trilateration (MT) procedure on the best observed solutions. To test the proposed algorithm, it is first compared with its two spin-offs (the proposed algorithm without the MT procedure and the proposed algorithm without the BL and MT procedures) and then compared with six existing optimization algorithms on ten randomly generated network topologies with four different connectivity ranges. The simulation results suggest that the proposed algorithm significantly outperforms the other algorithms in terms of estimating the positions of sensor nodes in WSNs. They also point out the effectiveness of applying the MT procedure and BL method to the proposed algorithm in solving the problem.