Bio-inspired algorithms for optimized deployment of sensor nodes in wireless networks

Abstract

Wireless Sensor Networks (WSNs) are a type of ad-hoc network technology that has been around for more than two decades. WSNs typically consist of a number of dedicated sensors, which are fundamentally low-cost, autonomous, resource-constrained devices organized into a cooperative network to perform a common monitoring task. From their first appearance until the present day, there has been a significant effort conducted by researchers to achieve a reliable WSN design that can provide better Quality of Service (QoS) for a wide range of applications. Deployment optimization is one of the crucial issues that must be taken into consideration while designing an efficient WSN. What is meant by deployment optimization is that the sensors must be placed in strategic locations that optimize one or multiple design criteria including, coverage, connectivity, cost, and network lifetime. In this thesis, we have addressed the problem of deployment by using bio-inspired algorithms. We proposed two deployment solutions for optimally placing homogeneous and heterogeneous WSNs. The proposed bio-inspired solutions allow the relocation of network sensors to locations that optimize coverage and energy efficiency. The first solution IBA achieves the desired objectives by eliminating both coverage redundancy and coverage holes resulted after the random deployment of heterogeneous sensors. Whereas the second solution BAGOA, which is developed by hybridizing two algorithms, namely the Bees Algorithm (BA) and the Grasshopper Optimization Algorithm (GOA), achieves high coverage and ensures low mobility during deployment in several deployment situations, even when the network is composed of both mobile and stationary sensors. The effectiveness of the proposed solutions is confirmed by the high performance recorded during the comparison with recently proposed solutions in the literature. The presented results show that IBA and BAGOA provide significantly better performance in all deployment test cases in terms of solution quality, convergence, and stability.