Metaheurísticas multiobjetivo para la resolución del problema del posicionamiento de nodos repetidores en redes de sensores inalámbricos

Abstract

A Wireless Sensor Network (WSN) is composed of a set of sensors, capturing information about the environment, and a sink node, which collects all the information provided by the network. WSNs are particularly sensitive to energy cost, especially with multi-hop routing protocols, where the devices send data to each other's. In recent years, a new device specialised in communication tasks and called Relay Node (RN) is added to traditional WSNs as a possible way to address this issue, resulting in the NP-hard optimisation Relay Node Placement Problem (RNPP). In this thesis, we tackle three different approaches of the RNPP, divided into two groups: outdoor and indoor networks. In the first approach, we study how to efficiently deploy energy-harvesting RNs in previously-established static outdoor WSNs for optimising average energy consumption and average coverage. The second approach is a more realistic version of the previous deployment problem, where we also optimise network reliability. Both approaches are solved by applying a wide range of Multi-Objective (MO) metaheuristics. Specifically, we implement NSGA-II, SPEA2, MO-VNS, MO-ABC, MOFA, MO-GSA, and MOEA/D algorithms. In the third approach and based on the knowledge obtained from the outdoor problem, we propose a new line of research not considered before in the literature: the deployment of low-cost static indoor WSNs, trying to leverage existing infrastructure. This new MO problem derives from the need to deploy low-cost networks for providing indoor localisation services, e.g. for domestic and industrial robots.