Thèses sur le sujet « Social networks – Mathematical models »

Thesis (MSc)--Stellenbosch University, 2011<br>ENGLISH ABSTRACT: With billions of current internet users interacting through social networks, the need has arisen to analyze the structure of these networks. Many authors have proposed random graph models for social networks in an attempt to understand and reproduce the dynamics that govern social network development. This thesis proposes a random graph model that generates social networks using a community-based approach, in which users’ affiliations to communities are explicitly modeled and then translated into a social network. Our approach explicitly models the tendency of communities to overlap, and also proposes a method for determining the probability of two users being connected based on their levels of commitment to the communities they both belong to. Previous community-based models do not incorporate community overlap, and assume mutual members of any community are automatically connected. We provide a method for fitting our model to real-world social networks and demonstrate the effectiveness of our approach in reproducing real-world social network characteristics by investigating its fit on two data sets of current online social networks. The results verify that our proposed model is promising: it is the first community-based model that can accurately reproduce a variety of important social network characteristics, namely average separation, clustering, degree distribution, transitivity and network densification, simultaneously.<br>AFRIKAANSE OPSOMMING: Met biljoene huidige internet-gebruikers wat deesdae met behulp van aanlyn sosiale netwerke kommunikeer, het die analise van hierdie netwerke in die navorsingsgemeenskap toegeneem. Navorsers het al verskeie toevalsgrafiekmodelle vir sosiale netwerke voorgestel in ’n poging om die dinamika van die ontwikkeling van dié netwerke beter te verstaan en te dupliseer. In hierdie tesis word ’n nuwe toevalsgrafiekmodel vir sosiale netwerke voorgestel wat ’n gemeenskapsgebaseerde benadering volg, deurdat gebruikers se verbintenisse aan gemeenskappe eksplisiet gemodelleer word, en dié gemeenskapsmodel dan in ’n sosiale netwerk omskep word. Ons metode modelleer uitdruklik die geneigdheid van gemeenskappe om te oorvleuel, en verskaf ’n metode waardeur die waarskynlikheid van vriendskap tussen twee gebruikers bepaal kan word, op grond van hulle toewyding aan hulle wedersydse gemeenskappe. Vorige modelle inkorporeer nie gemeenskapsoorvleueling nie, en aanvaar ook dat alle lede van dieselfde gemeenskap vriende sal wees. Ons verskaf ’n metode om ons model se parameters te pas op sosiale netwerk datastelle en vertoon die vermoë van ons model om eienskappe van sosiale netwerke te dupliseer. Die resultate van ons model lyk belowend: dit is die eerste gemeenskapsgebaseerde model wat gelyktydig ’n belangrike verskeidenheid van sosiale netwerk eienskappe, naamlik gemiddelde skeidingsafstand, samedromming, graadverdeling, transitiwiteit en netwerksverdigting, akkuraat kan weerspieël.

Epidemiologists rely on human interaction networks for determining states and dynamics of disease propagations in populations. However, such networks are empirical snapshots of the past. It will greatly benefit if human interaction networks are statistically predicted and dynamically created while an epidemic is in progress. We develop an application framework for the generation of human interaction networks and running epidemiological processes utilizing research on human mobility patterns and agent-based modeling. The interaction networks are dynamically constructed by incorporating different types of Random Walks and human rules of engagements. We explore the characteristics of the created network and compare them with the known theoretical and empirical graphs. The dependencies of epidemic dynamics and their outcomes on patterns and parameters of human motion and motives are encountered and presented through this research. This work specifically describes how the types and parameters of random walks define properties of generated graphs. We show that some configurations of the system of agents in random walk can produce network topologies with properties similar to small-world networks. Our goal is to find sets of mobility patterns that lead to empirical-like networks. The possibility of phase transitions in the graphs due to changes in the parameterization of agent walks is the focus of this research as this knowledge can lead to the possibility of disruptions to disease diffusions in populations. This research shall facilitate work of public health researchers to predict the magnitude of an epidemic and estimate resources required for mitigation.

The landscape of world migration involves multiple interacting movements of people at various geographic scales, posing significant challenges to the dyadic-independence assumption underlying standard migration models. To account for emerging patterns of multilateral migration relationships, we represent world migration as a time-evolving, spatial network. The nodes in the World Migration Network (WMN) are countries located in geographic space, and the edges represent migratory movements for each decade from 1960-2000. In the first part of the thesis, we characterise the spatial network structure of the WMN, with a particular focus on detecting and mapping mesoscopic structures called 'communities' (i.e., sets of countries with denser migration connections internally than to the rest of the WMN). We employ a method for community detection that simultaneously accounts for multilateral migration, spatial constraints, time-dependence, and directionality in the WMN. We then introduce an approach for characterising local (intracommunity) and global (intercommunity) connectivity in the WMN. On this basis, we define a threefold typology that distinguishes 'cave', 'bi-regional', and 'bridging' communities. These are characterised with distinct migration patterns, spatial network structures, and temporal dynamics: cave communities are tightly-knit enduring structures that channel local migration between contiguous countries; bi-regional communities merge migration between two distinct geographic regions; bridging communities have hub-and-spoke dynamic structures that emerge from globe-spanning movements. Our results suggest that the WMN is neither a globally interconnected network nor reproducing geographic boundaries but involves heterogeneous patterns of global and local ('glocal') migration connectivity. We examine a set of relational, homophily, and spatial mechanisms that could have possibly generated the 'glocal' structure we observe. We found that communities of different types arise from significantly different mechanisms. Our results suggest that migration communities can have important implications for world migration, as different types of community structure provide distinct opportunities and constraints, thereby distinctively shaping future migration patterns.

Multi Agent Systems (MAS) are computerised systems composed of autonomous software agents that interact to solve complex problems. Within a MAS, agents require some mechanism to coordinate their activities. In the MAS literature, norms have been widely used to coordinate agents’ activities. Thus, given a MAS, a major research challenge is how to synthesise a normative system, namely a collection of norms, which supports its agents’ coordination. This dissertation focuses on the automated synthesis of norms for open Multi- Agent Systems. In an open MAS, the agent population may change along time, agents may be developed by third parties and their behaviours are not known beforehand. These particular conditions make specially challenging to synthesise a normative system to govern an open MAS. The MAS literature has mainly investigated two general approaches to norm synthesis: off-line design, and on-line synthesis. The first approach aims at synthesising a normative system at design time. With this aim, it assumes that the MAS state space is known at design time and does not change at runtime. This goes against the nature of open MAS, and thus off-line design is not appropriate to synthesise their norms. Alternatively, on-line norm synthesis considers that norms are synthesised at runtime. Most on-line synthesis research has focused on norm emergence, which considers that agents synthesise their own norms, thus assuming that they have norm synthesis capabilities. Again, this cannot be assumed in open MAS. Against this background, this dissertation introduces a whole computational framework to perform on-line norm synthesis for open Multi-Agent Systems. Firstly, this framework provides a computational model to synthesise norms for a MAS at runtime. Such computational model requires neither knowledge about agents’ behaviours beforehand nor their participation in the norm synthesis pro- cess. Instead, it considers a regulatory entity that observes agents’ interactions at runtime, identifying situations that are undesirable for coordination to sub- sequently synthesise norms that regulate these situations. Our computational model has been conceived to be of general purpose so that it can be employed to synthesise norms in a wide range of application domains by providing little domain-dependent information. Secondly, our framework provides an abstract architecture to implement such regulatory entity (the so-called Norm Synthesis Machine), which observes a MAS and executes a synthesis strategy to synthe- sise norms. Thirdly, our framework encompasses a family of norm synthesis strategies intended to be executed by the Norm Synthesis Machine. Overall, this family of strategies supports multi-objective on-line norm synthesis Our first synthesis strategy, the so-called base, aims at synthesising effective normative systems that successfully avoid situations that are undesirable for a MAS’ coordination. Then, two further strategies (called iron and simon) go beyond effectiveness and also consider compactness as a norm synthesis goal. iron and simon take alternative approaches to synthesise compact normative systems that, in addition to effectively achieve coordination, are as synthetic as possible. This allows them to reduce agents’ computational efforts when reasoning about norms. A fourth strategy, the so-called lion, goes beyond effectiveness and compactness to also consider liberality as a synthesis goal. lion aims at synthesising normative systems that are effective and compact while preserving agents’ freedom to the greatest possible extent. Our final strategy is desmon, which is capable of synthesising norms by considering different degrees of reactivity. desmon allows to adjust the amount of information that is required to decide whether a norm must be included in a normative system or not. Thus, desmon can synthesise norms either by being reactive (i.e., by considering little information), or by being more deliberative (by employing more information). We provide empirical evaluations of our norm synthesis strategies in two application domains: a road traffic domain, and an on-line community domain. In this former domain, we employ these strategies to synthesise effective, compact and liberal normative systems that successfully avoid collisions between cars. In the latter domain, our strategies synthesise normative systems based on users’ complaints about inappropriate contents. In this way, our strategies implement a regulatory approach that synthesises norms when there is enough user consensus about the need for norms. Overall, this thesis advances in the state of the art in norm synthesis by providing a novel computational model, an abstract architecture and a family of strategies for on-line norm synthesis for open Multi-Agent Systems.<br>Els sistemes Multi-Agent (MAS) són sistemes computeritzats composats d’agents autònoms que interaccionen per resoldre problemes complexos. A un MAS, els agents requereixen algun mecanisme per a coordinar les seves activitats. A la literatura en Sistemes Multi-Agent, les normes han estat àmpliament utilitzades per coordinar les activitats dels agents. Per tant, donat un MAS, un dels majors reptes d’investigació és el de sintetizar el sistema normatiu, és a dir, la col·lecció de normes, que suporti la coordinació dels agents. Aquesta tesi es centra en la síntesi automàtica de normes per sistemes Multi-Agent oberts. A un MAS obert, la població d’agents pot canviar amb el temps, els agents poden ésser desenvolupats per terceres parts, i els comportaments dels agents són desconeguts per endavant. Aquestes condicions particulars fan especialment complicat sintetizar el sistema normatiu que reguli un sistema Multi-Agent obert. En general, la literatura en Sistemes Multi-Agent ha investigat dues aproximacions a la síntesi de normes: disseny off-line, i síntesi on-line. La primera aproximació consisteix a sintetizar un sistema normatiu en temps de disseny. Amb aquest propòsit, aquesta aproximació assumeix que l’espai d’estats d’un MAS és conegut en temps de disseny i no canvia en temps d’execució. Això va contra la natura dels sistemes Multi-Agent oberts, i per tant el disseny off-line no és apropiat per a sintetitzar les seves normes. Com a alternativa, la síntesi on-line considera que les normes són sintetizades en temps d’execució. La majoria de recerca en síntesi on-line s’ha centrat en la emergència de normes, que considera que els agents sintetizen les seves pròpies normes, per tant assumint que tenen la capacitat de sintetitzar-les. Aquestes condicions tampoc no es poden assumir en un MAS obert. Donat això, aquesta tesi introdueix un marc computacional per la síntesi on-line de normes en sistemes Multi-Agent oberts. Primer, aquest marc proveeix un model computacional per sintetizar normes per un MAS en temps d’execució. Aquest model computacional no requereix ni coneixement sobre els comportaments dels agents per endavant ni la seva participación en la síntesi de normes. En canvi, considera que una entitat reguladora observa les interaccions dels agents en temps d’execució, identificant situacions indesitjades per la coordinació i sintetizant normes que regulen aquestes situacions. El nostre model computacional ha estat dissenyat per a ésser de propòsit general per tal que pugui ser utilitzat a la síntesi de normes en un ampli ventall de dominis d’aplicació proporcionant només información clau sobre el domini. Segon, el nostre marc proveeix una arquitectura abstracta per implementar aquesta entitat reguladora, anomenada Màquina de Síntesi, que observa un MAS en temps d’execució i executa una estratègia de síntesi que s’encarrega de sintetizar normes. Tercer, el nostre marc incorpora una familia d’estratègies de síntesi destinades a ésser executades per una màquina de síntesi. En general, aquesta familia d’estratègies soporta la síntesi multi-objectiu i on-line de normes. La nostra primera estratègia, anomenada BASE, està dissenyada per sintetitzar sistemes normatius eficaços que evitin de manera satisfactòria situacions indesitjades per la coordinació d’un sistema Multi-Agent. Després, dues estratègies de síntesi, anomenades IRON i SIMON, van més enllà de la eficàcia i també consideren la compacitat com a objectiu de síntesi. IRON i SIMON prenen aproximacions alternatives a la síntesi de sistemes normatius compactes que, a més d’aconseguir la coordinació de manera efectiva, siguin tant sintètics com fos possible. Això permet a aquestes estratègies reduir els esforços computacionals dels agents a l’hora de raonar sobre les normes. Una quarta estratègia, anomenada LION, va més enllà de la eficàcia i la compacitat per considerar també la liberalitat com a objectiu de síntesi. Lion sintetitza sistemes normatius que són eficaços i compactes mentre preserven la llibertat dels agents tant com sigui possible. La nostra última estratègia és desmon, que és capaç de sintetizar normes considerant diferents graus de reactivitat. desmon permet ajustar la quantitat d’informació necessària per decidir si una norma cal que sigui o no inclosa a un sistema normatiu. DESMON pot sintetizar normes essent reactiu (considerant poca informació), o essent més deliberatiu (considerant més informació). En aquesta tesi presentem avaluacions empíriques de les nostres estratègies de síntesi en dos dominis d’aplicació: el domini del tràfic, i el domini de les comunitats on-line. En aquest primer domini, utilitzem les nostres estratègies per a sintetizar sistemes normatius eficaços, compactes i liberals que eviten colisions entre cotxes. Al segon domini, les nostres estratègies sintetizen sistemes normatius basant-se en les queixes dels usuaris de la comunitat sobre continguts inapropiats. D’aquesta manera, les nostres estratègies implementen un mecanisme de regulació que sintetiza normes quan hi ha suficient consens entre els usuaris sobre la necessitat de normes. Aquesta tesi avança en l’estat de l’art en síntesi de normes al proporcionar un novedós model computacional, una arquitectura abstracta i una familia d’estratègies per la síntesi on-line de normes per sistemes Multi-Agent oberts.

A cross-linking model for the activation of the A cell or immune accessory cell as a function of certain extracellular conditions is developed to determine the valency of the specific factor receptor on the A cell surface. It is found that such a determination can be made based on the FWHM of cross-linking curves which differ by a full order of magnitude between the bivalent receptor case and the monovalent receptor case. This determination can be made provided one can obtain accurate values for the equilibrium constants which characterize the system and provided that activation and IL-1 secretion is a linear function of cross-linking. It is also found that a determination of valence can be made if the equilibrium constants are such that substantial one receptor bridge formation takes place (one antibody molecule bound on both ends by the same receptor). This one-receptor bridge formation only takes place if the receptor is bivalent, and it presents itself in the cross-linking curve in a very distinctive manner. A second network model described as an ecological competition model of steady state lymphocyte populations is presented. This model, known as the symmetrical network theory is analysed numerically by integration of the differential equations and shown to provide a reasonable qualitative picture of the immune system's stable steady states, and offer a glimpse of state switching.<br>Science, Faculty of<br>Physics and Astronomy, Department of<br>Graduate

In the continuum modeling of traffic networks, a macroscopic cost-flow function (MCF) and macroscopic fundamental diagram (MFD) can be used to represent the fundamental relationships between traffic quantities such as speed, flow, and density. The MCF governs the steady-state cost-flow relationship, whereas the MFD represents the instantaneous inter-relationship between speed, flow, and density of traffic streams. This thesis explores the influence of network topologies on the MCF and MFD. The Hong Kong road system is divided into unit-sized road networks with various physical characteristics for which the network structure and signal timings are reserved. By universally scaling the origin-destination (OD) matrices of the morning peak, traffic conditions ranging from free-flow to congestion are created for microscopic simulation. From the simulation results, an MCF that relates the average journey time and the number of vehicles traveling through the network in one hour and an MFD that relates space to the mean speed and average density aggregated across 300s intervals are derived. The MCF and MFD are calibrated with mathematical models for each network. The density of roads, junctions, and signal junctions all influence the value of the macroscopic parameters in the MCF and MFD, and predictive equations are constructed that relate the macroscopic parameters to the network topological characteristics. Based on the fitting performance of the mathematical models, recommendations are made for selecting MCF and MFD models for continuum modeling.<br>published_or_final_version<br>Civil Engineering<br>Master<br>Master of Philosophy

Thesis (MSc)--Stellenbosch University, 2013.<br>ENGLISH ABSTRACT: Designing a mechanistic model that can give rise to realistic architecture of ecological networks is central to the understanding of how species assemble and function in ecosystems. As species are constantly adjusting their diets in an antagonistic network, we here incorporate this adaptive behaviour of diet choice into a bipartite network model, with the effect of antagonistic interactions between species depicted by Holling’s type II functional response. Predictions of this model fit extremely well with the observed levels of nestedness, modularity and node-degree distributions for 61 real host-parasitoid and plant-herbivore networks. We further examined two specific scenarios of our model (species with identical [neutral] demographic parameters and interactions with identical [neutral] benefit in the network) and found that the demography-neutral scenario overestimated observed modularity, whilst the benefit-neutral scenario over-estimate observed nestedness. Relationships between nestedness, modularity and connectance were found strong. Moreover, in contrast to the common belief of the high modularity in antagonistic networks, most real networks (> 80%) are significantly nested, whilst nearly 40% of the real networks are surprisingly less compartmentalized than random networks generated from null models. Regardless of the controversy on whether antagonistic networks are nested or compartmentalized, the proposed model captured the essence of the dynamic nature of structural emergence in antagonistic networks. Due to its predictive power, this model was further used to investigate robustness in antagonistic networks. Predictions showed that the robustness of a network is determined by many factors, such as connectance, resource degree distribution, resource-consumer ratio, diversity, nestedness and compartmentalisation. Surprisingly, the manner of network response to species loss was independent of the sequence followed while removing species from a network. Variations were only noticed in the intensity of the effect resulting from the removals. In addition, we also showed that species extinction procedures which ignore the interaction switch underestimate the effect of any loss of species in these networks. We must therefore value our knowledge of possible adaptive processes in the ecosystem as they may be important for resolving the diversity-stability debate.<br>AFRIKAANSE OPSOMMING: Die ontwerp van ’n meganistiese model wat aanleiding kan gee tot realistiese argitektuur van ekologiese netwerke is sentraal tot die begrip van hoe spesies bymekaar kom en funksioneer in ekosisteme. Soos spesies voortdurend hul dieet aanpas in ’n antagonistiese netwerk, het ons hierdie aanpasbare gedrag van dieet keuse in ’n bipartiet netwerk model ingewerk, met die effek van antagonistiese interaksies tussen spesies wat uitgebeeld word deur Holling se tipe II funksionele reaksie. Voorspellings van hierdie model pas baie goed met die waargenome vlakke van nestedness, modulariteit en node-graad uitkerings vir 61 ware gasheer-parasiet en plant-herbivoor netwerke. Verder het ons twee spesifieke gevalle van ons model (spesies met identiese [neutrale] demografiese parameters en interaksies met identiese [neutrale] voordeel in die netwerk) ondersoek en gevind dat die demografie-neutrale geval waargenome modulariteit oorskat, terwyl die voordeelneutraal geval waargenome nestedness oorskat. Verhoudings tussen nestedness, modulariteit en konnektiwiteit is sterk bevind. Verder, in teenstelling met die algemene verwagting van hoe modulariteit in antagonistiese netwerke, is oorhoofse werklike netwerke (> 80%) aansienlik geneste, terwyl byna 40% van die werklike netwerke is verbasend minder gekompartimenteerd as ewekansige netwerke gegenereer uit null modelle. Ongeag van die omstredenheid oor of antagonistiese netwerke geneste of gekompartimenteerd is, die voorgestelde model vang die essensie van die dinamiese aard van die strukturele opkoms in antagonistiese netwerke. As gevolg van sy voorspellende krag, is hierdie model verder gebruik om robuustheid te ondersoek in antagonistiese netwerke. Voorspellings het getoon dat die robuustheid van ’n netwerk word bepaal deur verskeie faktore, soos konnektiwiteit, hulpbron-graad verspreiding, hulpbron-verbruiker verhouding, diversiteit, nestedness en kompartementasie. Verrassend, die wyse van die netwerk reaksie op die verlies van spesies was onafhanklik van die reeks wat gevolg het toe die spesies verwyder is uit ’n netwerk. Variasies is slegs opgemerk in die intensiteit van die effek van die verskuiwings. Benewens, ons het ook aangetoon dat die prosedures van spesies se uitsterwing wat die interaksie skakelaar geignoreer het, onderskat die effek van ’n verlies van spesies in hierdie netwerke. Ons moet dus die waarde van ons kennis van die moontlike aanpassing prosesse in die ekosisteem in agneem, aangesien dit belangrik kan wees vir die oplossing van die diversiteit-stabiliteit debat.

In this thesis we examine mathematical models for dynamical reconfiguration of telecommunication networks. Dynamical configuration is one of the most important problems in the field of network management. In the case of some part of the network being damaged the system should maintain its operations, taking into account the new conditions. This can be achieved by using the residual capacity of the system if it is available. If there were not enough residual capacity, the managers of the system need to involve some of the unaffected traffic in order to reorganize traffic. A prolonged breakdown will push some customers to change their provider; there is also a lost [sic] of profit because of breakdowns, that can be considered as a certain implicit penalty. To reorganize traffic by using new routes we should have a flexible routing system.<br>Doctor of Philosophy

Il terrorismo di stampo jihadista rappresenta una minaccia per la società e una sfida per gli scienziati interessati a comprenderne la complessità. Questa complessità richiede costantemente nuovi sviluppi in termini di ricerca sul terrorismo. Migliorare la conoscenza empirica rispetto a tale fenomeno può potenzialmente contribuire a sviluppare applicazioni concrete e, in ultima istanza, a prevenire danni all’uomo. In considerazione di tali aspetti, questa tesi presenta un nuovo quadro metodologico che integra scienza delle reti, modelli stocastici e apprendimento profondo per far luce sul terrorismo jihadista sia a livello esplicativo che predittivo. In particolare, questo lavoro compara e analizza le organizzazioni jihadiste più attive a livello mondiale (ovvero lo Stato Islamico, i Talebani, Al Qaeda, Boko Haram e Al Shabaab) per studiarne i pattern comportamentali e predirne le future azioni. Attraverso un impianto teorico che si poggia sulla concentrazione spaziale del crimine e sulle prospettive strategiche del comportamento terroristico, questa tesi persegue tre obiettivi collegati utilizzando altrettante tecniche ibride. In primo luogo, verrà esplorata la complessità operativa delle organizzazioni jihadiste attraverso l’analisi di matrici stocastiche di transizione e verrà presentato un nuovo coefficiente, denominato “Normalized Transition Similarity”, che misura la somiglianza fra paia di gruppi in termini di dinamiche operative. In secondo luogo, i processi stocastici di Hawkes aiuteranno a testare la presenza di meccanismi di dipendenza temporale all’interno delle più comuni sotto-sequenze strategiche di ciascun gruppo. Infine, il framework integrerà la meta-reti complesse e l’apprendimento profondo per classificare e prevedere i target a maggiore rischio di essere colpiti dalle organizzazioni jihadiste durante i loro futuri attacchi. Per quanto riguarda i risultati, le matrici stocastiche di transizione mostrano che i gruppi terroristici possiedono un ricco e complesso repertorio di combinazioni in termini di armi e obiettivi. Inoltre, i processi di Hawkes indicano la presenza di diffusa self-excitability nelle sequenze di eventi. Infine, i modelli predittivi che sfruttano la flessibilità delle serie temporali derivanti da grafi dinamici e le reti neurali Long Short-Term Memory forniscono risultati promettenti rispetto ai target più a rischio. Nel complesso, questo lavoro ambisce a dimostrare come connessioni astratte e nascoste fra eventi possano essere fondamentali nel rivelare le meccaniche del comportamento jihadista e come processi memory-like (ovvero molteplici comportamenti ricorrenti, interconnessi e non randomici) possano risultare estremamente utili nel comprendere le modalità attraverso cui tali organizzazioni operano.<br>Jihadist terrorism represents a global threat for societies and a challenge for scientists interested in understanding its complexity. This complexity continuously calls for developments in terrorism research. Enhancing the empirical knowledge on the phenomenon can potentially contribute to developing concrete real-world applications and, ultimately, to the prevention of societal damages. In light of these aspects, this work presents a novel methodological framework that integrates network science, mathematical modeling, and deep learning to shed light on jihadism, both at the explanatory and predictive levels. Specifically, this dissertation will compare and analyze the world's most active jihadist terrorist organizations (i.e. The Islamic State, the Taliban, Al Qaeda, Boko Haram, and Al Shabaab) to investigate their behavioral patterns and forecast their future actions. Building upon a theoretical framework that relies on the spatial concentration of terrorist violence and the strategic perspective of terrorist behavior, this dissertation will pursue three linked tasks, employing as many hybrid techniques. Firstly, explore the operational complexity of jihadist organizations using stochastic transition matrices and present Normalized Transition Similarity, a novel coefficient of pairwise similarity in terms of strategic behavior. Secondly, investigate the presence of time-dependent dynamics in attack sequences using Hawkes point processes. Thirdly, integrate complex meta-networks and deep learning to rank and forecast most probable future targets attacked by the jihadist groups. Concerning the results, stochastic transition matrices show that terrorist groups possess a complex repertoire of combinations in the use of weapons and targets. Furthermore, Hawkes models indicate the diffused presence of self-excitability in attack sequences. Finally, forecasting models that exploit the flexibility of graph-derived time series and Long Short-Term Memory networks provide promising results in terms of correct predictions of most likely terrorist targets. Overall, this research seeks to reveal how hidden abstract connections between events can be exploited to unveil jihadist mechanics and how memory-like processes (i.e. multiple non-random parallel and interconnected recurrent behaviors) might illuminate the way in which these groups act.

The acquisition of a spatial structure during embryo development involves the differentiation of cells, often according to positional information. The complexity of the molecular networks regulating differentiation and of the mechanisms generating positional information makes it necessary to study them by means of mathematical modeling. Vertebrate embryos also acquire a segmented structure during somitogenesis this requires spatial and temporal variations in gene expression, which mathematical modeling can also help understand. A molecular mechanism for the somitogenesis clock is proposed, which accounts for inter-cellular synchronisation, and is based on positive feedback, even though it is compatible with all experimental data interpreted as showing that the clock is based on negative feedback. Experiments proposed to test this model involve real-time clock reporters, as well as inducible systems to induce spatially-controlled perturbations. Theoretical and experimental results have led to conflicting ideas as to how useful positional information can be established. In particular, it has been pointed out that some models of extracellular diffusion of morphogen exhibit inadequate traveling waves of receptor saturation. Two alternative (but not mutually exclusive) models are proposed, which are based on recent experimental results highlighting the roles of extracellular glycoproteins and morphogen oligomerization. The readout of positional information is translated to a discrete set of gene expression patterns. Intriguingly, it has been observed in numerous contexts that genes regulating differentiation are initially co-expressed in progenitors despite their antagonism. We characterise conditions under which three classes of generic "master regulatory networks" can behave as a "multi-switch", directing differentiation in an all-or-none fashion to a specific cell-type chosen among more than two possible outcomes. bHLH dimerisation networks can readily display coexistence of many antagonistic factors when competition is low. Decision-making can be forced by a transient increase in competition, which could correspond to some unexplained experimental observations related to Id proteins.

In this thesis I have analysed several mathematical models, which represent the dynamics of genetic regulatory networks. Methods of bifurcation analysis and direct numerical simulations were employed to study the biological phenomena that can occur due to the presence of time delays, such as stable periodic oscillations induced by Hopf bifurcations. To highlight the biological implications of time-delayed systems, different models of genetic regulatory networks as relevant to the onset and development of cancer were studied in detail, as well as genetic regulatory networks which describe the effects of transcription factors in the immune system. A network of an oscillator coupled with a switch was explored, as systems such as these are prevalent in genetic regulatory networks. The effects of time delays on its oscillatory and bistable behaviour were then investigated, the results of which were compared with available results from the literature.

A mathematical model has been developed which describes the fluid exchange from a capillary network of realistic topology, and calculates the spatial distribution of extravascular pressure. In this model, the capillaries are represented by a superposition of sources and sinks, resulting from a D'Arcy's Law description of flow in tissue of uniform fluid conductivity. The combination of this representation and Starling's Hypothesis, which relates the forces influencing transmural fluid exchange, yields an integral equation of the second kind which is solved numerically for the source strength distribution. Two important features of this approach are that: (i) it allows for interaction between the local tissue pressure field and fluid exchange (the model is called, therefore, the tissue pressure interaction model); and (ii) complex network morphologies are easily modeled. In single capillaries, this interaction, which decreases the predicted fluid exchange, increases with the magnitude of the ratio of capillary wall to extravascular fluid conductivities. For multiple capillaries, in addition to the "self" interaction of a capillary with the local extravascular pressure field, there is the possibility of interaction between capillaries ("capillary-capillary" interaction). The ratio of conductivities, and the additional factors of intercapillary distance and the number of capillaries, also affect interaction in capillary networks. Although interaction is only a weak function of intercapillary distance, it depends strongly on the number of capillaries. The major result from this work is that for the entire physiological range of conductivity ratios, interaction cannot be neglected in predicting fluid exchange. Although tissue pressure interaction affects the magnitude of fluid exchange, it does not greatly alter the pattern of extravascular flow. Therefore, previous models which neglected interaction are not invalidated by the present findings. The effect of interaction on planar capillary networks within a semi-infinite tissue space was also investigated. Flow boundary conditions were imposed at opposed planar boundaries, parallel with the capillary network. Interaction was found to decrease with decreasing distance between the boundary and plane of the capillaries. It still exerted a large effect, however, for distances greater than one-fourth the reference capillary length.

This research discusses a number of computer-based mathematical models which are designed to assist planners to make strategic decisions concerning the allocations of social services resources. A new model is postulated which uses current patterns of care to derive a set of alternative modes or packages of care, chooses a suitable set of allocations of clients to packages of care within given resource constraints and can be used to explore the effect on resource requirements of demographic changes, and to explore alternative ways of caring for clients if populations expand and/or resources are reduced. Comparisons are made with the DHSS Balance of Care model and with other models. An exploration is included of the weighting values used in the postulated model’s objective function.

Dans ce travail, nous présentons des techniques de réduction de données et de sécurité conçues pour économiser de l’énergie dans les réseaux de capteurs sans fil. Premièrement, nous proposons un modèle d’agrégation de données basé sur la fonction de similarité servant à éliminer les données redondantes. En plus, nous avons travaillé sur l’envoi, le moins possible, de caractéristiques de données en se basant sur des fonctions d’ajustement qui expriment ces caractéristiques. Deuxièmement, nous nous sommes intéressés à l’hétérogénéité des données tout en étudiant la corrélation entre ces caractéristiques multi variantes après avoir éliminé les mesures identiques durant la phase d’agrégation. Finalement, nous donnons un cadre de sécurité rigoureux, conçu à partir de la cryptographie, qui satisfait le niveau d’exigence atteint normalement dans les réseaux de capteurs sans fil arborescents. Il empêche les pirates d’obtenir des informations à propos des données détectées en assurant une certaine confidentialité de bout-en-bout entre les nœuds du capteur et le puits. Afin de valider nos techniques proposées, nous avons implémenté les simulations de la première technique sur des données collectées en temps réel à partir du réseau Sensor Scope déployé à Grand-St-Bernard. Les simulations de la deuxième et de la troisième technique sont réalisées sur des données collectées en temps réel à partir de 54 capteurs déployés au laboratoire de recherche Intel Berkeley. L’efficacité de nos techniques est évaluée selon le taux de réduction de données, la consommation de l’énergie, la précision des données et la complexité de temps<br>In this thesis, we present energy-efficient data reduction and security techniques dedicated for wireless sensor networks. First, we propose a data aggregation model based on the similarity function that helps in removing the redundant data. In addition, based on the fitting functions we worked on sending less data features, accompanied with the fitting function that expresses all features. Second, we focus on heterogeneity of the data while studying the correlation among these multivariate features in order to enhance the data prediction technique that is based on the polynomial function, all after removing the similar measures in the aggregation phase using the Euclidean distance. Finally, we provide a rigorous security framework inherited from cryptography satisfies the level of exigence usually attained in tree-based WSNs. It prevents attackers from gaining any information about sensed data, by ensuring an end-to-end privacy between sensor nodes and the sink. In order to validate our proposed techniques, we implemented the simulations of the first technique on real readings collected from a small Sensor Scope network which is deployed at the Grand-St-Bernard, while the simulations of the second and the third techniques are conducted on real data collected from 54 sensors deployed in the Intel Berkeley Research Lab. The performance of our techniques is evaluated according to data reduction rate, energy consumption, data accuracy and time complexity

Complex networks are composed of a large number of interacting nodes. Examples of complex networks include the topology of the Internet, connections between websites or web pages in the World Wide Web (WWW), and connections between participants in social networks.Due to their ubiquity, modeling complex networks is importantfor answering many research questions that cannot be answered without a mathematical model. For example, mathematical models of complex networks can be used to find the most vulnerable nodes to protect during a virus attack in theInternet, to predict connections between websites in the WWW, or to find members of different communities insocial networks. Researchers have analyzed complex networksand concluded that they are distinguished from other networks by four specific statistical properties. These four statistical properties are commonly known in this field as: (i) thesmall world effect,(ii) high average clustering coefficient, (iii) scale-free power law degree distribution, and (iv) emergence of community structure. These four statistical properties are further described later in this dissertation. Mostmodels used to generate complex networks attempt to produce networks with these statistical properties. Additionally, most of these network models generate homogeneous complex networks where all the networknodes are considered to have the same properties. Homogenous complex networks neglect the heterogeneous nature ofthe nodes in many complexnetworks. Moreover, somemodels proposed for generating heterogeneous complexnetworks are not general as they make specific assumptions about the properties of the network.Including heterogeneity in the connection algorithm of a modelwould makeitmore suitable for generating the subset of complex networks that exhibit selective linking.Additionally, all modelsproposed, to date, for generating heterogeneous complex networks do not preserve all four of the statistical properties of complexnetworks stated above. Thus, formulation of a model for the generation of general heterogeneous complex networkswith characteristics that resemble as much as possible the statistical properties common to the real-world networks that have received attention from the research community is still an open research question. In this work, we propose two new types of models to generate heterogeneous complex networks. First, we introduce the Integrated Attribute Similarity Model (IASM). IASM uses preferential attachment(PA) to connect nodes based on a similarity measure for node attributes combined with a node's structural popularity measure. IASM integrates the attribute similarity measure and a structural popularity measure in the computation of the connection function used to determine connectionsbetween each arriving (newly created) node and the existing(previously created or old) network nodes. IASM is also the first model known to assign an attribute vector having more than one element to each node, thus allowing different attributes per node in the generated complex network. Networks generated using IASM have a power law degree distribution and preserve the small world phenomenon. IASM models are enhanced to increase their clustering coefficient using a triad formation step (TFS). In a TFS, a node connects to the neighbor of the node to which it was previously connected through preferential attachment, thus forming a triad. The TFS increases the number of triads that are formed in the generated network which increases the network's average clustering coefficient. We also introduce a second novel model,the Settling Node Adaptive Model (SNAM). SNAM reflects the heterogeneous nature of connectionstandard requirements for nodes. The connectionstandard requirements for a noderefers to the values of attribute similarity and/or structural popularityof old node ythat node new xwould find acceptable in order to connect to node y.SNAM is novel in that such a node connection criterion is not included in any previous model for the generation of complex networks. SNAM is shown to be successful in preserving the power law degree distribution, the small world phenomenon, and the high clustering coefficient of complex networks. Next,we implement a modification to the IASM and SNAM models that results in the emergence of community structure.Nodes are classified into classes according to their attribute values. The connection algorithm is modified to include the class similarity values between network nodes. This community structure model preservesthe PL degree distribution, small world property, and does not affect average clustering coefficient values expected from both IASM and SNAM. Additionally, the model exhibits the presence of community structure having most of the connections made between nodes belonging to the same class with only a small percent of the connections made between nodes of different classes. We perform a mathematical analysis of IASM and SNAM to study the degree distribution for networks generated by both models. This mathematical analysis shows that networks generated by both models have a power law degree distribution. Finally, we completed a case study to illustrate the potential value of our research on the modeling of heterogeneous complex networks. This case study was performed on a Facebook dataset. The case study shows that SNAM, with some modifications to the connection algorithm, is capable of generating a network with almost the same characteristics as found for the original dataset. The case study providesinsight on how the flexibility of SNAM's connection algorithm can be an advantagethat makes SNAM capable of generating networks with different statistical properties. Ideas for future research areas includestudyingthe effect of using eigenvector centrality, instead of degree centrality, on the emergence of community structure in IASM; usingthe nodeindex as an indication for its order of arrival to the network and distributing added connections fairly among networknodes along the life of the generated network; experimenting with the nature of attributesto generatea more comprehensive model; and usingtime sensitive attributes in the models, where the attribute can change its value with time,<br>Ph. D.

Bibliography: pages 173-178.<br>Classical terrestrial (horizontal and vertical) networks and Doppler satellite derived networks are combined in a three dimensional transformation adjustment by solving for the external bias parameters using any of the three standard seven parameter similarity transformation models, namely the Bursa, Molodensky and Veis models. The object of this combination may be merely to merge the systems or networks, but may additionally involve an attempt to assign physical meaning to the estimated bias parameters. These two aspects, and the influence of the a priori Variance-Covariance matrix of the observables on the parameters and their interpretation is studied in detail. An in-depth conceptual, mathematical and numerical comparative assessment of the three standard models is made. The homogeneity of the classical terrestrial South African networks is investigated by comparing the transformation parameter sets derived for different regions and sub-regions of the country.

The use of switched linear networks is prevalent in power processing systems such as dc-to-dc power converters. These converters provide regulated dc output power and are widely used in the power supplies for computers and peripherals. As with most systems where high performance is a criterion, the power converter is usually enclosed in a feedback loop. Feedback design usually entails the use of small signal (linear) models with linear feedback design techniques. There exists a large number of small signal models, but most of them are ad hoc. A describing function approach has previously been presented which determines the exact small signal frequency response of the converters. This method is now used to systematically derive circuit models for power converters. The aim of the work presented in this thesis is to model the nonlinear part of the switching converters which are simply the switches. This study is focused on the Pulse Width Modulation (PWM) class of converters. Majority of available models are not very accurate because they are averaged models. The aim of the new approach is to obtain accurate and simple model. A partial realization of this aim has been achieved; the new model is accurate over a very wide frequency range, but presently the parameters are rather complicated.

Le train des spikes, la séquence des temps des potentiels d'action d'une cellule, est le donné habituellement analysé dans les enregistrementsélectrophysiologiques pour la description du pattern de décharge qui on suppose caractérisant les cellules neuronales. Nous présentons les résultats obtenus décrivant l'activité de décharge d'un petit réseau neuronal avec un modèle mathématique de saut-diffusion. Le potentiel membranaire du neurone en fonction du temps est donné par la somme d'un processus stochastique de diffusion et de deux processus de point qui provoquent des sauts d'amplitude constante à temps aléatoires discrètes. Différentes distributions sont considérées. Deux résultats principaux émergent. Le premier est que les histogrammes des intervalles entre potentiels d'action successifs montrent plus d'un maximum et un comportement de type resonnant. Ce fait suggère qu'en correspondance de chaque maximum la cellule ait une probabilité plus élevée de se décharger, de manière que le temps du potentiel d'action est un temps caractéristique de la cellule qui pourrait etre modulé en conditions physiologiques. Le deuxième résultat principal est que les afférences inhibitrices peuvent faciliter la transmission des potentiels d'action déchargés par l'unité afferente inhibitrice. Ce fait suggère que les cellules inhibitrices ne soient pas seulement impliquées en gardant équilibré l'excitabilité des toutes les neurones mais aussi dans le processus de l'information. La simulation de ce type de modèles exige une amélioration des algorithmes classiques. Dans ce cadre, la deuxième partie est dédiée àune étude purement théorique sur les processus bridge multidimensionnels<br>The spike train, i. E. The sequence of the action potential timings of a single unit, is the usual data that is analyzed in electrophysiological recordings for the description of the firing pattern which is supposed to characterize a certain type of cell. . We present the results obtained describing the firing activity of a small network of neurons with a mathematical jump diffusion model. That is the membrane potential as a function of time is given by the sum of a stochastic diffusion process and two counting processes that provoke jumps of constant sizes at discrete random times. Different distributions are considered for such processes. Two main results emerge. The first one is that interspike intervals (ISI) histograms show more than one peak (multimodality) and exhibit a resonant like behavior. This fact suggests that in correspondence of each mode (i. E. The lag of the maxima) the cell has a higher probability of firing such that the the lags become characteristic times of the cell which could be modulated under physiological conditions. The second main result concerns the role of inhibition in neuronal coding. Indeed we show that the inhibitory inputs may facilitate the transmission of the spikes generated by the excitatory inputs. This fact suggests that inhibitory cells are not only involved in keeping balanced the excitability of the cell but that they may also play a key role in the information process. Such kind of models requires to improve the algorithms to simulate the first passage time through a threshold of a stochastic process. So that the second part of this manuscript is dedicated to a a purely theoretical study on multidimensional bridge processes

In this thesis we present two different models of Complex Social Systems. The first model represents a vector-borne disease that takes place in a heterogeneous environment composed of areas of different types. Two populations take part in the epidemic process: humans and vectors. The population of humans moves around the heterogeneous environment. The idea of this model is to understand how the movement of people in the heterogeneous environment can affect the dynamics of the disease. The second model represents a Susceptible-Infected-Susceptible process on a social network. The population is represented as nodes, and the edges represent the possible transmissions between two people. We investigate how different topologies in the network affect the spread of the disease in the system. We simulate both models, and we perform a mathematical analysis of both of them. For the mathematical analysis we use an adapted version of the Random Heuristic Search framework, which was originally used for the understanding of Genetic Algorithms. In this thesis we investigate the predictability power of the mathematical approach.

Our daily online conversations with friends, family, colleagues, and strangers weave an intricate network of interactions. From these networked discussions emerge themes and topics that transcend the scope of any individual conversation. In turn, these themes direct the discourse of the network and continue to ebb and flow as the interactions between individuals shape the topics themselves. This rich loop between interpersonal conversations and overarching topics is a wonderful example of a complex system: the themes of a discussion are more than just the sum of its parts. Some of the most socially relevant topics emerging from these online conversations are those pertaining to racial justice issues. Since the shooting of Black teenager Michael Brown by White police officer Darren Wilson in Ferguson, Missouri, the protest hashtag #BlackLivesMatter has amplified critiques of extrajudicial shootings of Black Americans. In response to #BlackLivesMatter, other online users have adopted #AllLivesMatter, a counter-protest hashtag whose content argues that equal attention should be given to all lives regardless of race. Together these contentious hashtags each shape clashing narratives that echo previous civil rights battles and illustrate ongoing racial tension between police officers and Black Americans. These narratives have taken place on a massive scale with millions of online posts and articles debating the sentiments of "black lives matter" and "all lives matter." Since no one person could possibly read everything written in this debate, comprehensively understanding these conversations and their underlying networks requires us to leverage tools from data science, machine learning, and natural language processing. In Chapter 2, we utilize methodology from network science to measure to what extent #BlackLivesMatter and #AllLivesMatter are "slacktivist" movements, and the effect this has on the diversity of topics discussed within these hashtags. In Chapter 3, we precisely quantify the ways in which the discourse of #BlackLivesMatter and #AllLivesMatter diverge through the application of information-theoretic techniques, validating our results at the topic level from Chapter 2. These entropy-based approaches provide the foundation for powerful automated analysis of textual data, and we explore more generally how they can be used to construct a human-in-the-loop topic model in Chapter 4. Our work demonstrates that there is rich potential for weaving together social science domain knowledge with computational tools in the study of language, networks, and social movements.

An early event in embryo development is the formation of mesoderm, endoderm and ectoderm, known as the primary germ layers. The gene regulatory network (GRN) consisting of the regulatory mechanisms underlying the formation of mesoderm and endoderm (the mesendoderm GRN) has been extensively studied both experimentally and using mathematical models. The Xenopus GRN is complex, with much of this complexity due to large numbers of Mix and Nodal genes. Mice and humans have only single Mix and Nodal genes, meaning that the Xenopus GRN is overly complex compared with higher vertebrates. Urodele amphibians, for example the axolotl, have single Mix and Nodal genes required for mesoderm and endoderm formation giving a model organism for the study of a simplified mesendoderm GRN. We study the axolotl mesendoderm GRN by developing mathematical models that encompass the time evolution of transcription factors in a cell. A detailed investigation reveals that, despite differences in the axolotl mesendoderm GRN compared with the Xenopus, the model can qualitatively reproduce experimental observations. We obtain experimental data to estimate model parameters using a computational algorithm, then test the behaviour of the resulting mathematical model using independent data. We extend mathematical models of the Xenopus mesendoderm GRN to include transcription factors involved in patterning the DV axis. An investigation of this model shows that it can account for the formation of mesoderm, endoderm and anterior mesendoderm forming in regions of the embryo consistent wth experimental data. In the final section of this thesis, we extend a multicellular model of the Xenopus mesendoderm GRN into a grid of cells.

This dissertation concerns the problem of scheduling real-time traffic in wireless TDMA channels. The most important characteristic of real-time traffic is that it has straight end-to-end delay constraint. We begin the investigation by studying a scheduling principle which naturally achieves the best delay performance in stationary channel conditions. Since the resulting scheduling algorithm maintains equal flow delays across the whole system, it is termed the equal-delay policy. There are a number of advantages associated with this scheduling method. First, it is very simple and practical to implement in real system. Secondly, it can be easily modelled mathematically and admits an analytical solution, which is very important for the construction of an admission control algorithm, we present a mathematical model describing the dynamics of the scheduling system, as well as devising a tractable analytical solution to the problem. A third advantage of the equal-delay policy is that it can be easily extended to support flows with multiple delay constraints. We propose a multiple-class scheduling scheme based on similar allocation concepts as the equal-delay technique. The extended scheme can similarly be mathematically modelled and analytically characterized. A natural objection to the above proposed techniques is that wireless transmission resources can be under-utilized, since the scheduling algorithm pays no attention to the changing individual channel conditions. The reduction in channel utilization can also adversely affect the delay performance, We explain this phenomenon and study the impacts for a variety of different channel characteristics, Specifically, we propose an alternative channel-aware scheduling policy, which aims to maximize channel utilization while keeping a minimum probability of delay violation, The proposed channel-aware policy achieves near-optimal delay performance. However, unlike in the equal-delay case, the channel-aware policy is not practical to implement in a real system. The complicated system dynamics associated with the channel-aware scheme also hamper the development of a mathematical model and analytical solution for admission control. On the other hand, we observe from simulation results that under most circumstances, the equal-delay scheme achieves close to the pertonnance obtained by the channel-aware technique, With the additional benefits of simplicity and admitting analytical analysis. the equal-delay policy appears to be a more practical and suitable choice for scheduling real-time traffic in wireless networks.

Orientador: Ricardo de Oliveira Anido<br>Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Computação<br>Made available in DSpace on 2018-08-17T02:21:18Z (GMT). No. of bitstreams: 1 Pacheco_UrubatanRocha_M.pdf: 1174940 bytes, checksum: d2b5f4af6749eb4a1c7c6a1810b9749a (MD5) Previous issue date: 2010<br>Resumo: O foco deste trabalho é viabilizar a análise estrutural em redes sociais de colaboração científica a partir de bases de dados bibliográficos. Os dados bibliográficos são utilizados para obter redes sociais de afiliação dos autores a instituições de pesquisa científica, e das publicações são extraídas as suas relações com ontologias de áreas de pesquisa. Foram estudados e aplicados métodos que utilizam a análise das redes sociais para solução/redução de ambiguidades em identidades de nomes de pesquisadores, instituições, e veículos científicos. Outro assunto estudado foi a abordagem de medida da qualidade dos resultados e os problemas que afetam a sua qualidade. Concretizando o objetivo deste trabalho, foram construídas métricas e ferramentas que permitem a comparação da produção científica entre instituições, departamentos, áreas de pesquisa, países, etc. As ferramentas também produziram um ranking de universidades baseado no prestígio dos pesquisadores destas universidades na rede social de co-autoria. Este resultado permitiu demonstrar que a informação estrutural de prestígio foi devidamente capturada ao correlacionar este ranking com outros que avaliam a qualidade da produção científica das universidades utilizando critérios semelhantes.<br>Abstract: This work performs social network analysis of the scientific collaborations extracted from bibliographic data bases. The analysis also includes the authors' scientific institution afiliation, and its relation with the main scientific publications and with research subject ontologies. We studied and applied methods that use social network analysis to solve or mitigate the problem of ambiguity in researchers' identities. We also applied the methods for ambiguity resolution for names of institutions, scientific meeting venues, country/state names, etc. Another study subject was measuring the quality of the results. Finally we developed metrics and implemented tools that allow the comparison of the scientific production of institutions, researcher groups, research subjects fields, countries, etc. The tools also produced a ranking of universities based on the prestige of these universities researchers at the co-authorship social network. These results demonstrated that prestige structural information was properly captured showing its correlation with other works that assess the quality of scientific production of universities using similar criteria.<br>Mestrado<br>Metodologia e Tecnicas da Computação<br>Mestre em Ciência da Computação

Thesis (Ph. D.)--George Mason University, 2007.<br>Title from PDF t.p. (viewed Jan. 22, 2008). Thesis director: KC Chang. Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Information Technology. Vita: p. 117. Includes bibliographical references (p. 108-116). Also available in print.

The main goal of my studies has been to search for the optimal control strategy of controlling epidemics when taking into account both economical and social costs of the disease. Three control scenarios emerge with treating the whole population (global strategy, GS), treating a small number of individuals in a well-defined neighbourhood of a detected case (local strategy, LS) and allowing the disease to spread unchecked (null strategy, NS). The choice of the optimal strategy is governed mainly by a relative cost of palliative and preventive treatments. Although the properties of the pathogen might not be known in advance for emerging diseases, the prediction of the optimal strategy can be made based on economic analysis only. The details of the local strategy and in particular the size of the optimal treatment neighbourhood weakly depends on disease infectivity but strongly depends on other epidemiological factors (rate of occurring the symptoms, spontaneously recovery). The required extent of prevention is proportional to the size of the infection neighbourhood, but this relationship depends on time till detection and time till treatment in a non-nonlinear (power) law. The spontaneous recovery also affects the choice of the control strategy. I have extended my results to two contrasting and yet complementary models, in which individuals that have been through the disease can either be treated or not. Whether the removed individuals (i.e., those who have been through the disease but then spontaneously recover or die) are part of the treatment plan depends on the type of the disease agent. The key factor in choosing the right model is whether it is possible - and desirable - to distinguish such individuals from those who are susceptible. If the removed class is identified with dead individuals, the distinction is very clear. However, if the removal means recovery and immunity, it might not be possible to identify those who are immune. The models are similar in their epidemiological part, but differ in how the removed/recovered individuals are treated. The differences in models affect choice of the strategy only for very cheap treatment and slow spreading disease. However for the combinations of parameters that are important from the epidemiological perspective (high infectiousness and expensive treatment) the models give similar results. Moreover, even where the choice of the strategy is different, the total cost spent on controlling the epidemic is very similar for both models. Although regular and small-world networks capture some aspects of the structure of real networks of contacts between people, animals or plants, they do not include the effect of clustering noted in many real-life applications. The use of random clustered networks in epidemiological modelling takes an impor- tant step towards application of the modelling framework to realistic systems. Network topology and in particular clustering also affects the applicability of the control strategy.

Sistemas sociais têm se tornado cada vez mais relevantes para a Ciência da Computação em geral e para a Inteligência Artificial em particular. Tal interesse iniciou-se pela necessidade de analisar-se sistemas baseados em agentes onde a interação social destes agentes pode ter um impacto no resultado esperado. Uma tendência mais recente vem da área de Processamento Social de Informações, Computação Social e outros métodos crowdsourced, que são caracterizados por sistemas de computação compostos de pessoas reais, com um forte componente social na interação entre estas. O conjunto de todas interações sociais e os atores envolvidos compõem uma rede social, que pode ter uma forte influência em o quão eficaz ou eficiente o sistema pode ser. Nesta tese, exploramos o papel de estruturas de redes em sistemas sociais que visam a solução de problemas. Enquadramos a solução de problemas como uma busca por soluções válidas em um espaço de estados e propomos um modelo - a Rede Memética - que é capaz de realizar busca utilizando troca de informações (memes) entre atores interagindo em uma rede social. Tal modelo é aplicado a uma variedade de cenários e mostramos como a presença da rede social pode melhorar a capacidade do sistema em encontrar soluções. Adicionalmente, relacionamos propriedades específicas de diversas redes bem conhecidas ao comportamento observado para os algoritmos propostos, resultando em um conjunto de regras gerais que podem melhorar o desempenho de tais sistemas sociais. Por fim, mostramos que os algoritmos propostos são competitivos com técnicas tradicionais de busca heurística em diversos cenários.<br>Social systems are increasingly relevant to computer science in general and artificial intelligence in particular. Such interest was first sparkled by agent-based systems where the social interaction of such agents can be relevant to the outcome produced. A more recent trend comes from the general area of Social Information Processing, Social Computing and other crowdsourced systems, which are characterized by computing systems composed of people and strong social interactions between them. The set of all social interactions and actors compose a social network, which may have strong influence on how effective the system can be. In this thesis, we explore the role of network structure in social systems aiming at solving problems, focusing on numerical and combinatorial optimization. We frame problem solving as a search for valid solutions in a state space and propose a model - the Memetic Network - that is able to perform search by using the exchange of information, named memes, between actors interacting in a social network. Such model is applied to a variety of scenarios and we show that the presence of a social network greatly improves the system capacity to find good solutions. In addition, we relate specific properties of many well-known networks to the behavior displayed by the proposed algorithms, resulting in a set of general rules that may improve the performance of such social systems. Finally, we show that the proposed algorithms can be competitive with traditional heuristic search algorithms in a number of scenarios.

This thesis proposes new algorithms for restoration strategies that provision bandwidth guaranteed recovery for unicast and multicast connections. The primary focus is on online restoration strategies that sequentially do pre-planning of resource for each request using the current network resource state. Online restoration strategies do not require prior knowledge of all the requests like that of offline restoration strategies. Therefore, online restoration strategies are more suitable for on-demand and dynamic traffic engineering control. The proposed new algorithms are compared to known algorithms from literature. Most literature evaluates the performance of the algorithms with two metrics only: total bandwidth requirement and the number of requests accepted in the network. This thesis evaluates the algorithms in one additional dimension: the computational time. This is an important criterion when response times for establishing new connections are stringent. Each algorithm makes trade-off between computational complexity, bandwidth efficiency, and number of accepted requests. Results show that the proposed algorithms provide alternative trade-offs between the three performance metrics when compared to other existing algorithms. The alternatives provide more choice for the network providers and the best algorithm to use depends on the network's requirements. The restoration strategies used for unicast and multicast connections in this thesis are very compatible thus it is possible to integrate the restoration strategies into a single system where they share the same backup resources. Results from simulations show that using an integrated restoration model has significant benefits, which includes lower backup bandwidth requirement than the separate restoration model.

Chat rooms are of enormous interest to social network researchers as they are one of the most interactive internet areas. To understand the behaviour of users in a chat room, there have been studies on the analysis of the Response Waiting Time (RWT) based on traditional approaches of aggregating the network contacts. However, real social networks are dynamic and properties such as RWT change over time. Unfortunately, the traditional approach focuses only on static network and neglecting the temporal variation in RWT which may have lead to misrepresentation of the true nature of RWT. In order to determine the true nature of RWT, we analyse and compare the RWT of three online chat room logs (Walford, IRC and T-REX) putting into consideration the dynamic nature of RWT. Our research shows that the distribution of the RWT exhibits multi-scaling behaviour, which signi cantly a ects the current views on the nature of RWT. This is a shift from simple power-law distribution to a more complex pattern. The previous study on users RWT between pairs of people claims that the RWT has a power-law distribution with an exponent of 1. However, our research shows that multi-scaling behaviour and the exponent has a wider range of values which depend on the environment and time of day. The di erent exponents observed on di erent time scales suggest that the time context or environment has a signi cant in uence on users RWT. Furthermore, using the chat characterise, we predicted the factors which could minimize response waiting time and improving the friendship connection during online chat sessions. We apply our ndings to design an algorithm for chat thread detection. Here, we proposed two variations of cluster algorithm. The rst algorithm involves the traditional approach while in the second one, the temporal variations in RWT was taken into consideration to capture the dynamic nature of a text stream. An advantage of our proposed method over the previous models is that previous models have involved highly computationally intensive methods and often lead to deterioration in the accuracy of the result whereas our proposed approach uses a simple and effective sequential thread detection method, which is less computationally intensive.

This thesis presents the developed taxonomy of the security threats in agent-based distributed systems. Based on this taxonomy, a set of theories is developed to facilitate analyzng the security threats of the mobile-agent systems. We propose the idea of using the developed security risk graph to model the system's vulnerabilties.

In this thesis we present a new model for identifying dependencies within a gene regulatory cycle. The model incorporates both probabilistic and temporal aspects, but is kept deliberately simple to make it amenable for learning from the gene expression data of microarray experiments. A key simplifying feature in our model is the use of a compression function for collapsing multiple causes of gene expression into a single cause. This allows us to introduce a learning algorithm which avoids the over-fitting tendencies of models with many parameters. We have validated the learning algorithm on simulated data, and carried out experiments on real microarray data. In doing so, we have discovered novel, yet plausible, biological relationships.