Understanding the behavior of complex, large-scale, interconnected systems in a rigorous and structured manner is one of the most pressing scientific and technological challenges of current times. These systems include, among many others, transportation and communications systems, smart grids and power grids, financial markets etc. Failures of these systems have potentially enormous social, environmental and financial costs. In this work, we investigate the failure mechanisms of load-sharing complex systems. The systems are composed of multiple nodes or components whose failures are determined based on the interaction of their respective strengths and loads (or capacity and demand respectively) as well as the ability of a component to share its load with its neighbors when needed. Each component possesses a specific strength (capacity) and can be in one of three states: failed, damaged or functioning normally. The states are determined based on the load (demand) on the component. We focus on two distinct mechanisms to model the interaction between components strengths and loads. The first, a Loss of Strength (LOS) model and the second, a Customer Service (CS) model. We implement both models on lattice and scale-free graph network topologies. The failure mechanisms of these two models demonstrate temporal scaling phenomena, phase transitions and multiple distinct failure modes excited by extremal dynamics. We find that the resiliency of these models is sensitive to the underlying network topology. For critical ranges of parameters the models demonstrate power law and exponential failure patterns. We find that the failure mechanisms of these models have parallels to failure mechanisms of critical infrastructure systems such as congestion in transportation networks, cascading failure in electrical power grids, creep-rupture in composite structures, and draw-downs in financial markets. Based on the different variants of failure, strategies for mitigating and postponing failure in these critical infrastructure systems can be formulated.

Peer-to-peer networks and mobile ad hoc networks are emerging distributed networks that share several similarities. Fundamental among these similarities is the decentralized role of each participating node to route messages on behalf of other nodes, and thereby, collectively realizing communication between any pair of nodes. Messages are routed on a topology graph that is determined by the peer relationship between nodes. Although routing is fairly straightforward when the topology graph is static, dynamic variations in the peer relationship that often occur in peer-to-peer and mobile ad hoc networks present challenges to routing. In this thesis, we examine the interplay between routing messages and network topology design in two classes of these networks -- unstructured peer-to-peer networks and sparsely-connected mobile ad hoc networks. In unstructured peer-to-peer networks, we add structure to overlay topologies to support file sharing. Specifically, we investigate the advantages of designing overlay topologies with small-world properties to improve (a) search protocol performance and (b) network utilization. We show, using simulation, that "small-world-like" overlay topologies where every node has many close neighbors and few random neighbors exhibit high chances of locating files close to the source of file search query. This improvement in search protocol performance is achieved while decreasing the traffic load on the links in the underlying network. In the context of sparsely-connected mobile ad hoc networks where nodes provide connectivity via mobility, we present a protocol for routing in space and time where the message forwarding decision involves not only where to forward (space), but also when to forward (time). We introduce space-time routing tables and develop methods to compute these routing tables for those instances of ad hoc networks where node mobility is predictable over either a finite horizon or indefinitely due to periodicity in node motion. Furthermore, when the node mobility is unpredictable, we investigate several forwarding heuristics to address the scarcity in transmission opportunities in these sparsely-connected ad hoc networks. In particular, we present the advantages of fragmenting messages and augmenting them with erasure codes to improve the end-to-end message delivery performance.

This dissertation focuses on the material transport system design problem (MTSDP), integrating decisions of technology selection and flow network design. This research is motivated by the design of material transport systems (MTS) in manufacturing plants. The objective is to design a MTS with minimum lifetime costs, subject to service requirements, flow network restrictions, and limited resources. We characterize the MTSDP from the perspectives of task requirements, transport technology, and space utilization. A classification is proposed for transport technologies such that instances in the same class share the same properties, and a decision framework is proposed to emphasize the inter-relationships of three major decisions: task clustering, network connecting, and technology selection. We consider fixed and variable costs, arc capacities, and empty travel in our formulations. We propose two solution approaches for the MTSDP. The first is the compact formulation (CF) approach where the three major decisions are handled by a mixed integer non-linear programming (MINLP) formulation. Relaxation techniques are applied to linearize the model. The solution of the resulting linear formulation (MILP) provides a lower bound to that of MINLP. A tightened formulation reduces the computational time by a factor of 3.85. The experiment also shows that when control system costs are significant, designs with multiple-task clusters are more economical than those restricted to single-task clusters. The other approach is clustering/set partition (CSP), where the three decisions are decomposed and solved sequentially. In an example MTS design problem, three methods are compared: CSP, a GREEDY approach from the literature, and enumeration. CSP finds the optimal solution, while GREEDY results in 31% greater costs. A similar comparison with another example is made for the CF and CSP approaches. We apply the CSP approach in a case problem, using data from an auto parts manufacturer. We include flow path crossing constraints and perform experiments to determine solution quality over a range of small problem sizes. The largest difference from optimality is 3.34%, and the average is 0.98%. More importantly, based on these experiments, it seems there is no evidence that the difference percentage grows with an increase in the number of tasks.

Wireless sensor networks (WSN) are event based systems that rely on the collective effort of densely deployed sensor nodes continuously observing a physical phenomenon. The spatio-temporal correlation between the sensor observations and the cross-layer design advantages are significant and unique to the design of WSN. Due to the high density in the network topology, sensor observations are highly correlated in the space domain. Furthermore, the nature of the energy-radiating physical phenomenon constitutes the temporal correlation between each consecutive observation of a sensor node. This unique characteristic of WSN can be exploited through a cross-layer design of communication functionalities to improve energy efficiency of the network. In this thesis, several key elements are investigated to capture and exploit the correlation in the WSN for the realization of advanced efficient communication protocols. A theoretical framework is developed to capture the spatial and temporal correlations in WSN and to enable the development of efficient communication protocols. Based on this framework, spatial Correlation-based Collaborative Medium Access Control (CC-MAC) protocol is described, which exploits the spatial correlation in the WSN in order to achieve efficient medium access. Furthermore, the cross-layer module (XLM), which melts common protocol layer functionalities into a cross-layer module for resource-constrained sensor nodes, is developed. The cross-layer analysis of error control in WSN is then presented to enable a comprehensive comparison of error control schemes for WSN. Finally, the cross-layer packet size optimization framework is described.

Aujourd’hui les réseaux électriques sont exploités dans un marché dérégulé. Les gestionnaires des réseaux électriques sont tenus d’assurer un certain nombre de critères de fiabilité et de continuité du service, tout en minimisant le coût total consacré aux efforts effectués pour maintenir la fiabilité des installations. Il s’agit de trouver une stratégie, qui répond à plusieurs exigences, comme :le coût, les performances, la législation, les exigences du régulateur, etc. Cependant, le processus de prise de décision est subjectif, car chaque participant ramène sa contribution sur base de sa propre expérience. Bien que ce processus permette de trouver la « meilleure » stratégie, cette dernière n’est pas forcément la stratégie « optimale ». Ce compromis technico-économique a sensibilisé les gestionnaires des réseaux électriques à la nécessité d’un recours à des outils d’aide à la décision, qui doivent se baser sur des nouvelles approches quantitatives et une modélisation plus proches de la réalité physique.<p>Cette thèse rentre dans le cadre d’un projet de recherche lancé par ELIA, et dénommé COMPRIMa (Cost-Optimization Models for the Planning of the Renewal, Inspection, and Maintenance of Belgian power system facilities). Ce projet vise à développer une méthodologie qui permet de modéliser une partie du réseau électrique de transport (par les réseaux de Petri stochastiques) et de simuler son comportement dynamique sur un horizon donné (simulation de Monte Carlo). L’évaluation des indices de fiabilité permet de comparer les différents scénarios qui tentent d’améliorer les performances de l’installation. L’approche proposée est basée sur la stratégie RCM (Reliability-Centered Maintenance).<p>La méthodologie développée dans cette thèse permet une modélisation plus réaliste du réseau qui tient compte, entre autres, des aspects suivants :<p>- La corrélation quantitative entre le processus de maintenance et le processus de vieillissement des composants (par un modèle d’âge virtuel) ;<p>- Les dépendances liées à l’aspect multi-composant du système, qui tient compte des modes de défaillance spécifiques des systèmes de protection ;<p>- L’aspect économique lié à la stratégie de maintenance (inspection, entretien, réparation, remplacement), aux coupures (programmées et forcées) et aux événements à risque (refus disjoncteur, perte d’un client, perte d’un jeu de barres, perte d’une sous-station, etc.). / Doctorat en sciences appliquées / info:eu-repo/semantics/nonPublished