Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.741651
Title: Network structure and dynamics of empirical multiplex systems
Author: Stella, Massimo
ISNI:       0000 0004 7225 0834
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2017
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Abstract:
Interactions are important, since they can influence and shape a variety of real-world networked systems. Interactions can have a multi-relational nature, i.e. being of different type. Different layers of interactions can look very different from each other, motivating the need for developing multi-layer network models and metrics in network science. This thesis aims at developing novel multiplex frameworks for the quantitative investigation of two real-world systems: (i) the web of relationships between words in the human mind and (ii) the ecological interactions among animals in ecosystems. Despite being different in nature, both the mental lexicon of words and ecosystems can be represented as a multiplex network, where nodes represent distinct entities (e.g. words or animal groups) interacting on different layers in different ways (e.g. words being semantically and/or phonologically similar; animal species eating or parasitising each other). In both the considered systems, interactions crucially determine function and dynamics of a variety of processes. In the mental lexicon, individual interactions have been shown to influence both language acquisition and usage. In Part I of this thesis, I show that the structure of the phonological layer reflects constraints related to language use. I proceed by introducing the framework of multiplex lexical networks for quantifying, for the first time, the influence that phonology and semantics combined can have on (i) word acquisition of toddlers and (ii) word processing of adults. Results highlight phenomena that are not observable in single-layer networks. In toddlers word learning strategies based on the whole multiplex structure match empirical word learning significantly better than strategies based on individual layers, indicating that multiplexity is important for early word acquisition. At later ages the multiplex structure evolves by displaying an early, explosive emergence of a multiplex network core of words, which facilitates mental navigation and increases robustness against cognitive impairments. The second part of this thesis focuses on ecosystems, where interactions encapsulated in food webs or host-parasite networks greatly influence species extinction. In Part II of this thesis, I introduce the framework of ecological multiplex or “ecomultiplex” networks for combining predator-prey and host-parasite contact interactions as two layers of a network representing trophic links in a given ecosystem. I show that host-parasite interactions can dramatically increase the susceptibility of ecosystems to a parasite pandemic compared to models based on single-layer trophic networks only. Results of the ecomultiplex model are tested against empirical findings from field work in Brazilian ecosystems, finding agreement between my theoretical results and empirical data. Furthermore, by considering the multi-relational nature of trophic interactions, I quantitatively show that generalist top predators might accelerate parasite spread rather than hampering it, thus providing a theoretical explanation to recent empirical findings. Both in the mental lexicon and ecosystems, multiplexity influences structure, dynamics and function in ways not yet accounted for in the literature. This thesis aims to fill this gap by suggesting multiplex frameworks suitable for quantitative testing of empirical conjectures, while opening new modelling challenges at the interface of physics, network science and other disciplines.
Supervisor: Brede, Markus Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.741651  DOI: Not available
Share: