IoT networks 3D deployment using hybrid many-objective optimization algorithms

  • PDF / 2,507,573 Bytes
  • 47 Pages / 439.37 x 666.142 pts Page_size
  • 73 Downloads / 199 Views

DOWNLOAD

REPORT


IoT networks 3D deployment using hybrid many-objective optimization algorithms Sami Mnasri1 Thierry Val1

· Nejah Nasri2,3 · Malek Alrashidi2 · Adrien van den Bossche1 ·

Received: 17 November 2018 / Revised: 20 January 2020 / Accepted: 29 April 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract When resolving many-objective problems, multi-objective optimization algorithms encounter several difficulties degrading their performances. These difficulties may concern the exponential execution time, the effectiveness of the mutation and recombination operators or finding the tradeoff between diversity and convergence. In this paper, the issue of 3D redeploying in indoor the connected objects (or nodes) in the Internet of Things collection networks (formerly known as wireless sensor nodes) is investigated. The aim is to determine the ideal locations of the objects to be added to enhance an initial deployment while satisfying antagonist objectives and constraints. In this regard, a first proposed contribution aim to introduce an hybrid model that includes many-objective optimization algorithms relying on decomposition (MOEA/D, MOEA/DD) and reference points (Two_Arch2, NSGA-III) while using two strategies for introducing the preferences (PI-EMO-PC) and the dimensionality reduction (MVU-PCA). This hybridization aims to combine the algorithms advantages for resolving the many-objective issues. The second contribution concerns prototyping and deploying real connected objects which allows assessing the performance of the proposed hybrid scheme on a real world environment. The obtained experimental and numerical results show the efficiency of the suggested hybridization scheme against the original algorithms. Keywords IoT collection networks · 3D indoor redeployment · Experimental validation · Many-objective optimization · Preference incorporation · Dimensionality reduction

Electronic supplementary material The online version of this article (https://doi.org/10.1007/s10732-02 0-09445-x) contains supplementary material, which is available to authorized users. Extended author information available on the last page of the article

123

S. Mnasri et al.

1 Introduction To implement a wireless sensor network (WSN), the location of the nodes should be first chosen according to specific criteria in order to optimize several targeted objectives like coverage, localization, connectivity or consumption rate of energy. Thus, node deployment greatly influences the network performance and its operation. It aims essentially at proposing a network topology with well-defined number and positions of nodes. This deployment is said to be 3D if the variation of the heights between nodes are important with respect to the width and length of the “Region of Interest” (RoI). In this study, we investigate the 3D deployment which is more complicated and represents the RoI topography better than the 2D deployment. The migration of the WSNs to the Internet of Things (IoT) gave birth to the Internet of Things collection