Improved Many-Objective Optimization Algorithms for the 3D Indoor Deployment Problem

• Published in: Arabian Journal for Science and Engineering Vol. 44; no. 4; pp. 3883 - 3904
• Main Authors: Mnasri, Sami, Nasri, Nejah, van den Bossche, Adrien, Val, Thierry
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.04.2019; Springer Nature B.V; King Fahd University of Petroleum and Minerals SAUDI ARABIA - Springer (en ligne)
• Subjects: Algorithms; Computer Science; Engineering; Evolutionary algorithms; Humanities and Social Sciences; Mathematical models; multidisciplinary; Multiple objective analysis; Mutation; Networking and Internet Architecture; Nodes; Optimization; Remote sensors; Research Article - Computer Engineering and Computer Science; Science; Sensors; Wireless sensor networks; 3D indoor deployment; Experimental validation; Many-objective optimization; Adaptive operators; Neighbourhood
• ISSN: 2193-567X; 1319-8025; 2191-4281
• DOI: 10.1007/s13369-018-03712-7

Compared with the two-dimensional deployment, the three-dimensional deployment of sensor networks is more challenging. We studied the problem of 3D repositioning of sensor nodes in wireless sensor networks. We aim essentially to add a set of nodes to the initial architecture. The positions of the added nodes are determined by the proposed algorithms while optimizing a set of objectives. In this paper, we suggest two main contributions. The first one is an analysis contribution where the modelling of the problem is given and a set of modifications is incorporated on the tested multi-objective evolutionary algorithms to resolve the issues encountered when resolving many-objective problems. These modifications concern essentially an adaptive mutation and recombination operators with neighbourhood mating restrictions, the use of a multiple scalarizing functions concept and the incorporation of the reduction in dimensionality. The second contribution is an application one, where an experimental study on real testbeds is detailed to test the behaviour of the enhanced algorithms on a real-world context. Experimental tests followed by numerical results prove the efficiency of the proposed modifications against original algorithms.